AWS Startups Blog

AWS launches GenAI Loft Tour to accelerate AI innovation in startup and developer communities worldwide

Adam Seligman — Thu, 11 Jul 2024 13:01:10 +0000

Startups have always been the lifeblood of innovation. They are the frontrunners in new technologies adoption, and when it comes to generative artificial intelligence (AI), they are poised to transform industries and shape the future. That’s why we recently announced a commitment of $230 million to accelerate the creation of generative AI applications by startups around the world as well as the second-annual AWS Generative AI Accelerator. With these efforts, we are doubling down on our commitment to support startup founders to innovate faster and reinvent customer experiences and applications with generative AI.

Building on this news, today we are announcing the AWS GenAI Lofts, a global tour of pop-up collaborative spaces and immersive experiences for startups and developers that will take residence in innovation and AI hubs around the globe. This initiative also represents our commitment to make it easy for developers of all skill levels to build and scale generative AI applications. Similar to the AWS Startup Lofts that started in 2014, the AWS GenAI Lofts provide a one-stop destination for in-person engagement for startups and developers to learn how to use and implement generative AI technology, get up to speed on the latest trends, and connect with a wider community of technology and business experts.

With AWS GenAI Lofts, startups, developers, and AI enthusiasts can get hands-on AI products and services from AWS Partners and AWS, including Amazon Bedrock and Amazon Q. Developers will have the opportunity to gain greater understanding of advanced techniques, such as building agentic workflows and tuning foundation models, and dig deeper into generative AI use cases and demos. Visitors can experience exclusive sessions led by industry luminaries, make connections with generative AI investors and leaders, and get their questions answered in-person by generative AI experts.

Pop-ups will open for up to 12 weeks in global innovation hubs, including:

AWS GenAI Loft | Bengaluru: July 29, 2024
AWS GenAI Loft | San Francisco: August 12, 2024
AWS GenAI Loft | São Paulo: September 2, 2024
AWS GenAI Loft | London: September 30, 2024
AWS GenAI Loft | Paris: October 8, 2024

Visitors will benefit from immersive experiences showcasing cutting-edge generative AI projects, workshops, fireside chats, and hands-on programming from AI experts and AWS partners Anthropic, Cerebral Valley, and Weights & Biases, among others. Each city will host a roster of esteemed AI experts and thought leaders, including machine learning data scientists. These sessions will provide access to some of the brightest minds in the field with daily events. Visitors can also leverage the Ask an Expert Bar to get their questions answered by AWS Solutions Architects.

While each pop-up GenAI Loft is open, startups and developers can take advantage of programs, workshops, and tools, including:

Experiential Series

This series highlights not just luminary speakers, but incorporates seeing, feeling, and even touching generative AI through real-life applications. Experiences include seeing immersive robotics art-making in motion, spotlighting cultural influencers who are using generative AI to enrich experiences. These “Artist-in-Residence” sessions are interactive and experiential, bringing generative AI technology to life.

Bootcamps

Bootcamps are immersive three-day programs designed to equip founders and developers with new skills, frameworks, and resources to propel their business forward. The programs are tailored to the needs of both business and technical founders, providing a combination of hands-on workshops and interactive sessions.

Immersion Days

AWS Solution-Focused Immersion Days are a series of events designed to provide startups employees and developers with hands-on experience using generative AI services and discover efficient methodologies to help problem-solve through generative AI.

Startup Talks

Startup Talks provide a comprehensive learning experience for startup founders, covering technical, business, and personal aspects of the entrepreneurial journey, while also incorporating the perspectives of investors, industry partners, and startup experts. These talks will feature a mix of content led by leaders and founders from some of the top startups in the world.

To find out when the AWS GenAI Loft tour is coming to a city near you, get more details on programming, and register, visit aws.amazon.com/startups/lp/aws-gen-ai-lofts.

Accelerating the next wave of generative AI startups

Swami Sivasubramanian — Thu, 13 Jun 2024 12:17:43 +0000

Since day one, AWS has helped startups bring their ideas to life by democratizing access to the technology powering some of the largest enterprises around the world including Amazon. Each year since 2020, we have provided startups nearly $1 billion in AWS Promotional Credits. It’s no coincidence then that 80% of the world’s unicorns use AWS. I am lucky to have had a front row seat to the development of so many of these startups over my time at AWS—companies like Netflix, Wiz, and Airtasker. And I’m enthusiastic about the rapid pace at which startups are adopting generative artificial intelligence (AI) and how this technology is creating an entirely new generation of startups.

These generative AI startups have the ability to transform industries and shape the future, which is why today we announced a commitment of $230 million to accelerate the creation of generative AI applications by startups around the world. We are excited to collaborate with visionary startups, nurture their growth, and unlock new possibilities. In addition to this monetary investment, today we’re also announcing the second-annual AWS Generative AI Accelerator in partnership with NVIDIA. This global 10-week hybrid program is designed to propel the next wave of generative AI startups. This year, we’re expanding the program 4x to serve 80 startups globally. Selected participants will each receive up to $1 million in AWS Promotional Credits to fuel their development and scaling needs. The program also provides go-to-market support as well as business and technical mentorship. Participants will tap into a network that includes domain experts from AWS as well as key AWS partners such as NVIDIA, Meta, Mistral AI, and venture capital firms investing in generative AI.

Building in the cloud with generative AI

In addition to these programs, AWS is committed to making it possible for startups of all sizes and developers of all skill levels to build and scale generative AI applications with the most comprehensive set of capabilities across the three layers of the generative AI stack. At the bottom layer of the stack, we provide infrastructure to train large language models (LLMs) and foundation models (FMs) and produce inferences or predictions. This includes the best NVIDIA GPUs and GPU-optimized software, custom, machine learning (ML) chips including AWS Trainium and AWS Inferentia, as well as Amazon SageMaker, which greatly simplifies the ML development process. In the middle layer, Amazon Bedrock makes it easier for startups to build secure, customized, and responsible generative AI applications using LLMs and other FMs from leading AI companies. And at the top layer of the stack, we have Amazon Q, the most capable generative AI-powered assistant for accelerating software development and leveraging companies’ internal data.

Customers are innovating using technologies across the stack. For instance, during my time at the VivaTech conference in Paris last month, I sat down Michael Chen, VP of Strategic Alliances at PolyAI, which offers customized voice AI solutions for enterprises. PolyAI develops natural-sounding text-to-speech models using Amazon SageMaker. And they build on Amazon Bedrock to ensure responsible and ethical AI practices. They use Amazon Connect to integrate their voice AI into customer service operations.

At the bottom layer of the stack, NinjaTech uses Trainium and Inferentia2 chips, along with Amazon SageMaker, to build, train, and scale custom AI agents. From conducting research to scheduling meetings, these AI agents save time and money for NinjaTech’s users by bringing the power of generative AI into their everyday workflows. I recently sat down with Sam Naghshineh, Co-founder and CTO, to discuss how this approach enables them to save time and resources for their users.

Leonardo.AI, a startup from the 2023 AWS Generative AI Accelerator cohort, is also harnessing the capabilities of AWS Inferentia2 to enable artists and professionals to produce high-quality visual assets with unmatched speed and consistency. By reducing their inference costs without sacrificing performance, Leonardo.AI can offer their most advanced generative AI features at a more accessible price point.

Leading generative AI startups, including Perplexity, Hugging Face, AI21 Labs, Articul8, Luma AI, Hippocratic AI, Recursal AI, and DatologyAI are building, training, and deploying their models on Amazon SageMaker. For instance, Hugging Face used Amazon SageMaker HyperPod, a feature that accelerates training by up to 40%, to create new open-source FMs. The automated job recovery feature helps minimize disruptions during the FM training process, saving them hundreds of hours of training time a year.

At the middle layer, Perplexity leverages Amazon Bedrock with Anthropic Claude 3 to build their AI-powered search engine. Bedrock ensures robust data protection, ethical alignment through content filtering, and scalable deployment of Claude 3. While Nexxiot, an innovator in transportation and supply chain solutions, quickly moved its Scope AI assistant solution to Amazon Bedrock with Anthropic Claude in order to give their customers the best real-time, conversational insights into their transport assets.

At the top layer, Amazon Q Developer helps developers at startups build, test, and deploy applications faster and more efficiently, allowing them to focus their valuable energy on driving innovation. Ancileo, an insurance SaaS provider for insurers, re-insurers, brokers, and affinity partners, uses Amazon Q Developer to reduce the time to resolve coding-related issues by 30%, and is integrating ticketing and documentation with Amazon Q to speed up onboarding and allow anyone in the company to quickly find their answers. Amazon Q Business enables everyone at a startup to be more data-driven and make better, faster decisions using the organization’s collective knowledge. Brightcove, a leading provider of cloud video services, deployed Amazon Q Business to streamline their customer support workflow, allowing the team to expedite responses, provide more personalized service, and ultimately enhance the customer experience.

Resources for generative AI startups

The future of generative AI belongs to those who act now. The application window for the AWS Generative AI Accelerator program is open from June 13 to July 19, 2024, and we’ll be selecting a global cohort of the most promising generative AI startups. Don’t miss this unique chance to redefine what’s possible with generative AI, and apply now!

Other helpful resources include:

You can use your AWS Activate credits for Amazon Bedrock to experiment with FMs, along with a broad set of capabilities needed to build responsible generative AI applications with security and privacy.
Dive deeper by exploring our Generative AI Community space for technical content, insights, and connections with fellow builders. AWS also provides free training to help the current and future workforce take advantage of Amazon’s generative AI tools. For those interested in learning to build with generative AI on AWS, explore the comprehensive Generative AI Learning Plan for Developers to gain the skills you need to create cutting-edge applications
NVIDIA offers NVIDIA Inception, a free program designed to help startups evolve faster through cutting-edge technology, opportunities to connect with venture capitalists, and access to the latest technical resources from NVIDIA.

Apply now, explore the resources, and join the generative AI revolution with AWS.

Additional Resources

Twitch series: Let’s Ship It – with AWS! Generative AI

AWS Generative AI Accelerator Program: Apply now

AWS and CrowdStrike announce the winner of the AWS & CrowdStrike Cybersecurity Accelerator

AWS Editorial Team — Tue, 14 May 2024 18:24:33 +0000

Left to Right: CJ Moses, CISO, Amazon, Mike Sentonas, President, CrowdStrike, Galina Antova, Co-founder, Claroty, Tal Kollender, Co-founder & CEO, Gytpol, Ilan Leiferman, Head BD Cybersecurity, AWS, Daniel Bernard, CBO, CrowdStrike, Dona Haj, VCs & Startups BD, AWS

In a groundbreaking initiative aimed at nurturing the next wave of cybersecurity disruptors, AWS and CrowdStrike announced security remediation platform GYTPOL as the winner’s of the 2024 AWS and CrowdStrike Cybersecurity Startup Accelerator. The top prize was awarded for their innovative technology, which specializes in continuous detection and automated remediation of device misconfigurations, ensuring zero impact on business continuity. The winner was announced after competing against 8 other startups at the final stages of the program.

Handpicked from a pool of hundreds of applicants, 23 startups embarked on an exciting journey through the AWS and Crowdstrike Cybersecurity Accelerator, a 10-week, equity-free program that provided unparalleled access to industry experts, masterclasses, global investors, and up to $25,000 in AWS Activate Credits, empowering these startups to scale and innovate in the cybersecurity realm.

Participating in the AWS and Crowdstrike Accelerator Program has truly been a game-changer for GYTPOL. The invaluable content and strategy sessions have significantly boosted our company’s growth. What’s been most rewarding on a personal level is the exceptional mentorship and the opportunity to connect with industry leaders I’ve long admired. It’s surreal to witness their genuine interest and eagerness to help. Equally exciting is meeting the rising stars, individuals whose potential is palpable. This program has a knack for cultivating talent and fostering a supportive community. I feel privileged to be part of this cohort, surrounded by such incredible tech and talent, in a program that’s truly one-of-a-kind.

Tal Kollender, CEO of Gytpol

The winner was unveiled at an exclusive startup showcase at the San Francisco Mint during RSA Conference 2024, attended by cybersecurity executives, technology buyers, CISOs, and investors. Mike Sentonas, CrowdStrike President, CJ Moses, AWS Chief Information Security Officer, and entrepreneur and investor, Galina Antova.

“CrowdStrike took cybersecurity to the cloud by building on AWS – together we’ve propelled how companies build and secure their businesses in the cloud, from code to runtime,” shared Daniel Bernard, Chief Business Officer at CrowdStrike and showcase moderator. “Partnering with AWS to foster the next-generation of cybersecurity innovators underscores our commitment to driving industry transformation.”

“In the midst of every crisis there may be an opportunity. But in technology, the reverse is often true, as well: every opportunity also creates a crisis. AI is undoubtedly a generational opportunity—a force multiplier for productivity that has quickly become essential for teams to compete. But it’s also a force multiplier for malicious actors and represents a new cybersecurity attack vector,” said Bogomil Balkansky, partner at Sequoia Capital. “Our portfolio company, Apex, helps teams resolve the tension between leveraging transformative AI innovations and ensuring robust protection against risks. We are proud to see Apex recognized as a finalist in the AWS & CrowdStrike accelerator.”

“Congratulations to GYTPOL for their outstanding achievement in winning the AWS and CrowdStrike Cybersecurity Startup Accelerator. Their innovative approach to continuous detection and automated remediation of device misconfigurations is truly groundbreaking. We are thrilled to see them thrive and look forward to supporting their continued success as they make waves in the cybersecurity industry,” said Kellen O’Connor, EMEA Managing Director for Startups at AWS.

Meet the Finalists of the 2024 Cohort:

Aim Security

Aim Security’s mission is to empower security teams to allow enterprise adoption of generative AI technologies securely and safely. Aim builds the first generative AI security platform and provides protection, risk management and governance for all Large Language Model (LLM) risks.

Miggo

Miggo is the world’s first Application Detection and Response platform. Uniquely integrating real-time identity-aware application tracing, proactive threat hunting, and anomaly detection at the business logic layer, Miggo reduces risk, ensures compliance, and protects even insecure production applications and APIs.

Oligo Security

Oligo Security is on a mission to proactively protect software throughout the development lifecycle. With contextual detection of exploitable flaws in all application code, the Oligo platform helps organizations focus on risks that matter – improving transparency and trust for security and engineering teams. Let your developers focus on building features, not fixes.

Opus Security

Opus Security brings together vulnerabilities from Cloud, Application, and other attack surfaces, consolidating organizational security postures. The platform automates prioritization, streamlines remediation, and promotes collaboration between security, engineering, and IT. With Opus, organizations benefit from an efficient cross-organizational risk reduction process that works.

Apono

Apono is a cloud privileged access management platform providing visibility and control to privileged permissions. Apono helps companies reduce over-privileges, automate manual access provisioning, and gain control of access with “Just In Time” and “Just Enough” dynamic provisioning.

Mindflow

Mindflow is an AI-driven automation platform designed for SecOps that empowers enterprise teams to operate at a new level of performance by intuitively automating repetitive, mundane tasks and seamlessly orchestrating all their tools.

onum

With the market’s largest library of integrated services and revolutionary generative AI automation and onum, empowering organizations with tailored cybersecurity strategies.

Find more information on this year’s AWS & CrowdStrike Cybersecurity Accelerator and discover other accelerator program opportunities available to startups by visiting Startups.AWS.

Ilan Leiferman is the Global Head of BD Cybersecurity Startups and VCs, developing the strategy for how AWS engages and support the fastest-growing cyber startups globally. Prior to AWS, Ilan founded an innovation advisory firm, The Shelf, serving Fortune500 customers to partner-up with startups, and was a venture partner of Bright Pixel Capital and Axon Partners Fund of Funds. Ilan also co-founded a startup named Wiffinity and a digital marketing agency: Rubik DS. He has a BA in Business Management, and recently completed an INSEAD executive program on Leadership Development. He is married with two, Irish twin, boys 4 & 5.

Boosted.ai’s generative AI portfolio manager surfaces near-instant finance insights with AWS

Ryan Masciovecchio — Tue, 07 May 2024 14:11:38 +0000

By switching from a generic LLM that was too expansive and cumbersome for their needs to a model tailored to their domain (capital markets), Boosted.ai reduced costs by 90 percent, vastly improved efficiency, and unlocked the GPU capacity needed to scale their generative AI investment management application.

Summary

In 2020, Boosted.ai expanded their artificial intelligence (AI)-powered financial analysis platform—Boosted Insights—by building an AI portfolio assistant for asset managers on a large language model (LLM) that processed data from 150,000 sources. The output was macro insights and market trend analysis on over 60,000 stocks across every global equity market (North America, EU & UK, APAC, Middle East, Latin America, and India). But using an LLM came with some significant drawbacks—a high annual cost to operate and GPU capacity limitations that limited their ability to scale.

Boosted.ai began domain-optimizing a model running on AWS and:

reduced costs by 90 percent without sacrificing quality
moved from overnight to near real-time updates, unlocking more value for their investment manager clients acting on hundreds of thousands of data sources
improved security and personalization with the ability to run a model in a customer’s private cloud, rather than running workloads through an LLM cloud

Introduction

2023 was the year generative AI went mainstream. Enhancing efficiency to do more with less will continue to be on corporate agendas throughout 2024 and beyond. It is critical for teams to have a strategy for how they will incorporate generative AI to create productivity gains. However, even when there’s a clear use case, it’s not always apparent how to implement generative AI in a way that makes sense for a business’s bottom line.

Here’s how Boosted.ai incorporated generative AI to automate research tasks for their investment management clients in a way that improved outcomes for both Boosted.ai and their customers.

Founded in 2017, Boosted.ai offers an AI and machine learning (ML) platform—Boosted Insights—to help asset managers sort through data to enhance their efficiency, improve their portfolio metrics, and make better, data-driven decisions. When the founders saw the impact of powerful LLMs, they decided to use a closed-source LLM to build an AI-powered portfolio management assistant. Overnight, it would process millions of documents from 150,000 sources, including nontraditional datasets like SEC filings such as 10Ks and 10Qs, earnings calls, trade publications, international news, local news, even fashion. After all, if you’re talking about a company like Shein going public, a Vogue article could become relevant investing information. Boosted Insights summarized and collated all this information into an interactive user interface that their asset manager clients could sort through themselves.

With their new generative AI model, Boosted.ai was now pushing critical investment information to all their clients, over 180 of the world’s biggest asset managers. For these teams, time is money. When something impacts a company’s stock price, how fast someone gets and acts on that information can be the difference of thousands, even millions of dollars. Boosted.ai gave these managers an edge. For instance, it flagged that Apple was moving some of its manufacturing capabilities into India before news broke in mainstream media outlets, because Boosted Insights was reading articles in Indian media.

Adding a generative AI component to Boosted Insights automated a lot of the research to turn an investing hypothesis into an actual trade. For instance, if an investor was concerned about a trade war with China, they could ask Boosted Insights: “What are the kinds of stocks I should buy or sell?” Before generative AI, answering that question was a 40-hour research process, sifting through hundreds of pages of analyst reports, news articles, and earnings summaries. With an AI-powered portfolio management assistant, 80 percent of that work was now automated.

Figure 1. Boosted Insights maps the performance of stocks with exposure to generative AI

Solving for scale with domain-specific language models

Boosted.ai’s generative AI rollout was extremely well received by clients, but the company wanted to scale it to run up to 5x or 10x more analysis and get from overnight reports to a true real-time system. But there was a problem: running the AI cost nearly $1 million a year in fees, and even if they wanted to buy more GPU capacity, they simply couldn’t. There just wasn’t enough GPU capacity for their AI financial analysis tool to scale into a real-time application.

Right-sizing the model for lower costs and greater scale

Boosted.ai’s challenges are increasingly common ones for organizations adopting LLMs and generative AI. Since LLMs are trained for general purpose use, the companies that train these models spend a lot of time, testing, and money to get them to work. The larger the model, the more accelerated compute it has to use on every request. As a result, for most organizations, including Boosted.ai, it is just not viable to use an LLM for a specific task.

Boosted.ai decided to explore a more targeted and cost-effective approach: fine-tuning a smaller language model to perform a specific task. In the AI/ML world, these models are often referred to as “open source,” but that doesn’t mean they are hacked together by random people sharing a wiki, as you might imagine from the early days of open-source coding. Instead, open-source language models, like Meta’s Llama 2, are trained on trillions of data points and maintained in secure environments like Amazon Bedrock. The difference is an open-source model gives users total access to its parameters and the option to fine-tune them for specific tasks. Closed-source LLMs, by contrast, are a black box that don’t allow for the kind of customization Boosted.ai needed to create.

The ability to fine-tune their model would prove to make a difference for Boosted.ai. Through the AWS Partner Network, Boosted.ai connected with Invisible, whose global network of AI training specialists allowed Boosted.ai to stay focused on their core developmental work while Invisible provided high-quality data annotation faster and more cost effectively than staffing an in-house team to the project. Together, AWS, Invisible, and Boosted.ai found and implemented the smallest possible model that could handle their use case, benchmarking against the industry-standard Massive Multitask Language Understanding (MMLU) dataset to evaluate performance.

Our goal was to have the smallest possible model with the highest possible IQ for our tasks. We went into the MMLU and looked at subtasks we thought were highly relevant to what Boosted.ai is doing: microeconomics and macroeconomics, math, and a few others. We grabbed the smallest model we thought would work and tuned it to be the best it could be for our tasks. If that didn’t work, we moved to the next size model and the next level of intelligence. — Joshua Pantony, Boosted.ai co-founder and CEO

With a more compact and efficient model that performed just as well at financial analysis, Boosted.ai slashed costs by 90 percent. The big benefit they saw from this efficiency was being able to massively upsize the amount of data they pulled—going from overnight updates to near real-time. More importantly, they got the GPUs they needed to scale. Where Boosted.ai once needed A100 and H100 to run their models, this more efficient domain-specific generative AI allowed them to run a layer on smaller and more readily available hardware.

Figure 2. Benefits of using a domain-specific model instead of a closed-source LLM

Better security and customization with a smaller model

Having fine-tuned a smaller model with the same efficacy, Boosted.ai had the computational capacity to run even more analysis. Now instead of processing data overnight, they could process data every minute and promise customers a delay of only 5-10 minutes between something happening and Boosted Insights picking it up.

The model also gave Boosted.ai more optionality for where and how they deploy. With an LLM, Boosted.ai was shipping the workload out to a closed-source cloud, getting the results back, and then storing it. Now, they can deploy inside another customer’s virtual private cloud (VPC) on AWS for added security.

Having a generative AI strategy will be a fundamental expectation for investment management firms in 2024, and we are seeing huge demand of companies wanting to run their internal data through our generative AI to create smart agents. Understandably, leveraging proprietary data raises privacy concerns. A lot of our users feel safer on our model than on a big closed-source LLM. 90 percent of our clients have an AWS account, and the benefit we’re seeing is that keeping their data secure within their private AWS cloud is extremely simple when we run on the same cloud.

Giving access to private deployments running their data is a lot easier than trying to build the entire thing from scratch. — Joshua Pantony

With the extra peace of mind that a private endpoint offers, more customers are willing to share their proprietary data to create more customized insights. For instance, a hedge fund might have access to interviews with hundreds of CFOs and management analysts. That dataset is too valuable and confidential to send to a public API endpoint. With Boosted.ai’s domain-specific approach, it doesn’t have to. The entire workload runs within the customer’s cloud, and they get more customized insights.

The future: domain-specific language models and a new way to tap expertise

As Boosted.ai’s fine-tuned smaller language model grows, the insights it offers will get crisper and more quantified. For instance, today it can say which companies are affected by an event, like the war in Ukraine. In the future, it will be able to quantify that effect and say, “exactly 7 percent of this company’s revenue will be impacted, and here’s the probability of how it will be impacted.”

Additionally, obtaining those insights will require less user interaction. It will be possible to upload your expertise and knowledge to your personalized AI, have it scan a vast database of information, and push unique ideas to you.

AI is the most rapidly adopted technology in human history, and for smaller organizations, today’s cutting-edge use cases are likely to be table stakes in a few short years.

We’re in this really unique time in history where there’s a lot of big companies that don’t know the potential of this technology and are adopting it in suboptimal ways. You’re seeing a ton of chatbots go up left, right, and center. If you’re a startup today, meet customers, learn their problems, and be aware of what generative AI is capable of. If you do, there’s a very high probability you’re going to find a unique value add.

Once you’re confident that you’ve got some product-market fit, I would think about fine-tuning smaller models versus LLMs across speed, accuracy, and data sensitivity. If you think any of those are critical for your use case, it’s probably worth it to use a domain-specific model. — Joshua Pantony

Additional thanks to Invisible for their contributions to this project and article. Invisible is an operations innovation company that seamlessly merges AI and automation with a skilled human workforce to unlock strategic execution bottlenecks.

How Patronus AI helps enterprises boost their confidence in generative AI

Aditya Shahani — Thu, 02 May 2024 14:21:31 +0000

Over the past few years, and especially since the launch of ChatGPT in 2022, the transformational potential of generative artificial intelligence (AI) has become undeniable for organizations of all sizes and across a wide range of industries. The next wave of adoption has already begun, with companies rushing to adopt generative AI tools in order to drive efficiency and enhance customer experiences. A 2023 McKinsey report estimated that generative AI could add the equivalent of $2.6 trillion to $4.4 trillion in value to the global economy annually, boosting AI’s overall economic impact by some 15-40 percent, while IBM’s latest CEO survey found that 50 percent of respondents were already integrating generative AI into their products and services.

As generative AI goes mainstream, however, customers and businesses are increasingly expressing concern about its trustworthiness and reliability. And it can be unclear why given inputs lead to certain outputs, making it difficult for companies to evaluate the results of their generative AI. Patronus AI, a company founded by machine learning (ML) experts Anand Kannappan and Rebecca Qian, has set out to tackle this problem. With its AI-driven automated evaluation and security platform, Patronus helps its customers use large language models (LLMs) confidently and responsibly while minimizing the risk of errors. The startup’s aim is to make AI models more trustworthy and more usable. “That’s become the big question in the past year. Every enterprise wants to use language models, but they’re concerned about the risks and even just the reliability of how they work, especially for their very specific use cases,” explains Anand. “Our mission is to boost enterprise confidence in generative AI.”

Reaping the benefits and managing the risks of generative AI

Generative AI is a type of AI that uses ML to generate new data similar to the data it was trained on. By learning the patterns and structure of the input datasets, generative AI produces original content—images, text, and even lines of code. Generative AI applications are powered by ML models that have been pre-trained on vast amounts of data, most notably LLMs trained on trillions of words across a range of natural language tasks.

The potential business benefits are sky-high. Firms have shown interest in using LLMs to leverage their own internal data through retrieval, to produce memos and presentations, to improve automated chat assistance, and to auto-complete code generation in software development. Anand also points to the whole range of other use cases that have not yet been realized. “There’s a lot of different industries that generative AI hasn’t disrupted yet. We’re really just at the early innings of everything that we’re seeing so far.”

As organizations consider expanding their use of generative AI, the issue of trustworthiness becomes more pressing. Users want to ensure their outputs comply with company regulations and policies while avoiding unsafe or illegal outcomes. “For larger companies and enterprises, especially in regulated industries,” explains Anand, “there are a lot of mission-critical scenarios where they want to use generative AI, but they’re concerned that if a mistake happens, it puts their reputation at risk, or even their own customers at risk.”

Patronus helps customers manage these risks and boost confidence in generative AI by improving the ability to measure, analyze, and experiment with the performance of the models in question. “It’s really about making sure that, regardless of the way that your system was developed, the overall testing and evaluation of everything is very robust and standardized,” says Anand. “And that’s really what’s missing right now: everyone wants to use language models, but there’s no really established or standardized framework of how to properly test them in a much more scientific way.”

Enhancing trustworthiness and performance

The automated Patronus platform allows customers to evaluate and compare the performance of different LLMs in real-world scenarios, thereby reducing the risk of undesired outputs. Patronus uses novel ML techniques to help customers automatically generate adversarial test suites and score and benchmark language model performance based on Patronus’s proprietary taxonomy of criteria. For example, the FinanceBench dataset is the industry’s first benchmark for LLM performance on financial questions.

“Everything we do at Patronus is very focused around helping companies be able to catch language model mistakes in a much more scalable and automated way,” says Anand. Many large companies are currently spending vast amounts on internal quality assurance teams and external consultants, who manually create test cases and grade their LLM outputs in spreadsheets, but Patronus’s AI-driven approach saves the need for such a slow and expensive process.

“Natural Language Processing (NLP) is quite empirical, so there is a lot of experimentation work that we are doing to ultimately figure out which evaluation techniques work the best,” explains Anand. “How can we enable those kinds of things in our product so that people can leverage the value … from the techniques that we figured out work the best, very easily and quickly? And how can they get performance improvements, not only for their own system, but even for the evaluation against that system that they’ve been able to do now because of Patronus?”

What results is a virtuous cycle: the more a company uses the product and gives feedback via the thumbs or thumbs down feature, the better its evaluations become, and the better the company’s own systems become as a result.

Boosting confidence through improved results and understandability

To unlock the potential of generative AI, improving its reliability and trustworthiness is vital. Potential adopters across a variety of industries and use cases are regularly held back—not just by the fact that mistakes are sometimes made by AI applications—but also by the difficulty of understanding how or why a problem has occurred, and how to avoid that happening in the future.

“What everyone is really asking for is a better way to have a lot more confidence in something when you roll it out to production,” says Anand. “And when you put it in front of your own employees, and even end customers, then that’s hundreds, thousands, or tens of thousands of people, so you want to make sure that those kinds of challenges are limited as much as possible. And, for the ones that do happen, you want to know when they happen and why.”

One of Patronus’ key goals is enhancing the understandability, or explainability, of generative AI models. This refers to the ability to pinpoint why certain outputs from LLMs are the way they are, and how customers can gain more control over those outputs’ reliability.

Patronus incorporates features aimed at explainability, primarily by giving customers direct insight into why a particular test case passed or failed. Per Anand: “That’s something that we do with natural language explanations, and our customers have told us that they liked that, because it gives them some quick insight into what might have been the reason why things have failed—and maybe even suggestions for improvements on how they can iterate on the prompt or generation parameter values, or even for fine-tuning … Our explainability is very focused around the actual evaluation itself.”

Looking toward the future of generative AI with AWS

To build their cloud-based application, Patronus has worked with AWS since the beginning. Patronus uses a range of different cloud-based services; Amazon Simple Queue Service (Amazon SQS) for queue infrastructure and Amazon Elastic Compute Cloud (Amazon EC2) for Kubernetes environments, they take advantage of the customization and flexibility available from Amazon Elastic Kubernetes Service (Amazon EKS).

Having worked with AWS for many years before he helped found Patronus, Anand and his team were able to leverage their familiarity and experience with AWS to quickly develop their product and infrastructure. Patronus has also worked closely with AWS’s startup-focused solutions teams, which have been “instrumental” in setting up connections and conversations. “The customer-focused aspect [at AWS] is always great, and we never take that for granted,” says Anand.

Patronus is now looking optimistically forward, having been inundated with interest and demand in the wake of its recent launch from stealth mode with $3 million in seed funding led by Lightspeed Venture Par tners. The team has also recently announced the first benchmark for LLM performance on financial questions, something co-designed with 15 financial industry domain experts.

“We are really excited for what we’re going to be able to do in the future,” says Anand. “And we’re going to continue to be focused on AI evaluation and testing, so being able to help companies identify gaps in language models…and understand how they can quantify performance, and ultimately get better products that they can build a lot more confidence around in the future.”

Ready to unleash the benefits of generative AI with confidence and reliability? Visit the AWS Generative AI Innovation Center for guidance planning, execution support, Generative AI use cases—or any other solution of your choice.

Technology that teaches empathy? How mpathic uses AI to help us listen to each other

Bonnie McClure — Wed, 01 May 2024 04:10:34 +0000

On a basic human level, we want to be heard. We want to connect with others, and we want to be understood. Unfortunately, we’re often faced with many things competing for our attention, which makes us bad listeners.

Active listening is a learned behavior and not easy to master. But what if artificial intelligence (AI) could augment our ability to really listen and truly relate to others? What if technology could draw upon our collective lived experiences and help us be more human to each other?

These are the questions Dr. Grin Lord, clinical psychologist and founder of conversation analytics company mpathic, has spent the last 15 years chasing. During her research, Grin and the team at mpathic have identified trust-building words, phrases, and communication behaviors and modeled them using AI.

“We look at what is promoting trust, what is promoting engagement, and how those impact outcomes,” explains mpathic’s Chief Innovation Officer, Dr. Danielle Schlosser.

In pursuit of a technology-driven approach to unlock empathy, mpathic developed something unique: a solution that not only analyzes and assesses the health of conversations but also provides recommendations for increasing their levels of empathy, trust, and engagement in real-time.

“Our differentiator is trying to be more behavioral and actionable,” says Grin. “We want to coach people on how to improve.”

Drawing on responses from a diverse range of experts with extensive empathy training, mpathic’s API quickly tags instances of misunderstanding within ongoing conversations and immediately offers feedback and suggestions on how to listen and respond with more empathy.

The results have been astonishing. When deployed in clinical trials, healthcare providers using mpathic’s API have been seven times more likely to capture participant risk and provide critical feedback. Similarly, in sales and HR software as a service (SaaS) use cases, businesses using mpathic products witnessed more customer engagement, satisfaction, and other outcomes.

Iterating on empathy education

Taking context and nuance into consideration, mpathic defines empathy as “accurate understanding.” But designing a successful method for teaching empathy turned out to be much more elusive than defining it.

In the early 2000s, Grin began her journey as part of a research study working with drivers involved in drunk driving accidents. The experiment consisted of brief interventions, including 15 minutes of empathic listening, showing acceptance and understanding of the driver’s experience. This brief empathic intervention led to reductions in drinking that held over three years later and a 46 percent reduction in readmissions to the hospital.

After that, Grin trained medical professionals on how to listen with empathy, teaching behaviors such as reflective listening, asking open-ended questions instead of closed-ended ones, and using affirmations.

When she found that a two-day workshop was not enough time to change deep-seated behaviors and styles of communication, she retooled her approach. Grin learned techniques from a nationwide phone coaching study where doctors would record themselves giving feedback. A psychologist would listen and provide doctors with performance-based suggestions on how to improve. This process could take weeks, so in 2008 she seized an opportunity to use machine learning (ML) to speed up the process.

At the University of Washington, Grin was a part of the team that built the first speech signal processing pipelines for performance-based feedback in a medical settings. “With the computing power at the time, it took about 6 hours to process a 30-minute call,” she says. “But the fact you could get any feedback the same day was considered really revolutionary.”

Now, with enhanced computing, power the original vision of performance-based feedback for medical providers was accelerated to actual real-time. Over the years, Grin built a team of dedicated subject matter experts and specialists pulling from those involved in the original research at University of Washington, as well as AI experts at Carnegie Mellon University, and industry experts from big tech.

The idea for mpathic came about when Grin and team realized the commercial value of empathic listening: “Could we make an API that would instantly take any communication and make it more empathic, regardless of the use case?”

The team built some of mpathic’s first models using data collected from Empathy Rocks, an empathy training game. In the game, therapists, including members of the Idaho State Crisis Line and California Indian Health Service, would respond to anonymous users from data in public forums with empathy and rank each other’s statements; they received continuing education for playing these games. “We had really diverse groups of people building these models through crowdsourcing that information,” explains Grin.

Expanding empathy training and tools across industries

As mpathic continues to evolve and grow their capabilities, the startup now has more than 200 different models for communication behaviors with tips and suggestions, including how to improve collaboration and power-sharing, and listen with more accuracy using reflections and open-ended questions. They also measure more unconscious metrics of human alignment, like language style synchrony, that have been found in Grin’s research to be more predictive of objective ratings of empathy than other skills. “The goal is not to replace human experience,” says Dr. Amber Jolley-Paige, Vice President of Clinical Product, “but to enhance it.”

With a tailored and flexible approach, mpathic uses analysis and metrics to support customers’ specific needs and KPIs, whatever the industry. They currently offer a suite of AI-powered products: the core mpathic API, mConsult, and mTrial. The core API integrates into other software, analyzing communications and proposing actionable suggestions. For example, when mpathic used their API to analyze recruitment interviews for different companies, they found that those who received empathetic feedback had an 8 percent increase in candidate acceptance. mConsult provides immediate recommendations and coaching by reviewing audio or video recordings. And mTrial streamlines clinical trials by enhancing data quality and ensuring consistent care, while proactively reducing risk and easing medical professionals’ workloads.

Envisioning the future of health equity

mpathic’s journey shows no signs of slowing down. To better reach their goal of improving human communication, the team is expanding its API to specifically address diverse cultural behaviors and coach providers in cultural adaptation.

Culture can affect how people communicate in various ways. For example, it may affect communication styles, how people deliver information, and their attitudes toward conflict. “With mpathic, we have the ability like never before to create more empathy in healthcare interactions and imagine a future where we can leverage AI to improve health equity,” says Dr. Alison Cerezo, Head of Research and Health Equity.

The startup built training data from a diverse group of different genders, cultures, and backgrounds to help curb AI bias. “A lot of the issues that you see with AI bias comes down to models built from data collected from only one or two backgrounds and not understanding the lived experience of the people that those models will impact,” explains Grin. mpathic ensures that they regularly build, refine, and deploy their models with attention and alignment to an ethical AI framework.

Moving forward, the team at mpathic plans to continue developing AI tools that recognize the nuanced and diverse viewpoints present in all human interactions. “There is no limit to the potential of this technology to train anyone to listen with empathy,” says Grin.

Going big with AWS

To scale their platform, mpathic needed a robust infrastructure. AWS provided a reliable, solid foundation for mpathic to grow and innovate securely. “We built on AWS to help us scale effectively and meet our customers’ needs quickly and seamlessly,” says Grin. “We’re a relatively tiny startup to be serving customers globally. To be able to tell our customers that we can host data wherever they are in the world is awesome, and wouldn’t be possible without AWS.” mpathic uses AWS for all of its foundational platform components, including compute, storage, and networking infrastructure, ensuring secure cross-border data transfer and storage.

Beyond technology, collaboration between mpathic and AWS was built on a shared commitment to helping mpathic reach their goals. “There is a degree of interest and support that’s really impressive, especially coming from such a large organization,” says Danielle. “It’s not just about the technology, it’s also about the connections.”

“AWS has also done a lot of work highlighting women founders, which I think is great,” adds Megan Greenlaw, Vice President of Life Sciences and Psychedelic AI. “To me it signifies a shift that’s happening in venture, the fact that a company can raise over $10 million and that 90% of those checks are being written by women is pretty outstanding,” says Grin.

Saving the Earth in 60 Seconds, One Startup at a Time

Chalaire Miller — Tue, 30 Apr 2024 16:56:18 +0000

As Earth Month draws to a close, we’re thrilled to showcase the innovative startups that participated in our “Saving the Earth in 60 Seconds, One Startup at a Time” challenge. These pioneering companies are each driving sustainability forward in their respective industries, from carbon planning to all-electric rideshare services and more. Let’s take a closer look at the startups featured!

1. Clarasight

“At Clarasight, we believe that for companies to achieve their emissions goals, they need to know more than just what has happened in the past. They need to be able to look into the future to see what will happen ahead.”

Adam Braun, CEO and co-founder of Clarasight, is leading the charge in forward-looking carbon planning and analysis software. With a focus on analytics, forecasting, and agile planning, Clarasight empowers organizations to seamlessly align their emissions goals—and their financial goals. Learn how Clarasight is using real-time data insights and integration capabilities to revolutionize sustainable decision-making, helping companies close the gap between intention and action.

See Adam Braun, CEO and co-founder of Clarasight make his pitch to see if he can make his case in just 60 seconds.

2. Revel

“Today in New York City, we have the four largest charging depots in the entire metro area; we run an all-electric rideshare fleet of over 500 vehicles. The entire rideshare sector in the city is going all-electric by 2030, by city mandate. Here at Revel, we are doing everything we can to accelerate that.”

Frank Reig, co-founder and CEO of Revel, is transforming urban mobility with electric vehicles (EVs). Revel’s public, fast-charging network and all-electric rideshare service are driving dense urban cities toward cleaner, greener futures. With expertise in technology, engineering, and clean energy, Revel is accelerating the adoption of EVs and reshaping urban landscapes, starting with the most densely populated urban area in the United States: New York City.

Our final founder in the ‘Saving the Earth in 60 Seconds, One Startup at a Time’ hashtag#EarthMonth challenge is Frank Reig, co-founder and CEO of Revel, an electric mobility and infrastructure company with a mission to accelerate EV adoption in America’s densest cities.

Watch Frank Reig, co-founder and CEO of Revel, talk about transforming urban mobility with EVs.

3. BlocPower

“Buildings in America represent 30 percent of our total US emissions. There’s no path to addressing the climate crisis without going building to building and greening all the buildings.”

Donnel Baird, CEO and founder of BlocPower, has set out to green every city in America by 2030. BlocPower utilizes proprietary technology to upgrade homes and buildings with clean, energy-efficient, electric technology. By streamlining processes and making upgrades accessible, BlocPower is redefining sustainability in the built environment. Learn how BlocPower has already used artificial intelligence (AI) to build digital models of all 125 million buildings in America—and is currently working closely with four American cities to implement this plan, electrifying and decarbonizing every one of their buildings.

Learn how Donnel Baird, CEO and founder of BlocPower, is revolutionizing buildings for smarter, greener, healthier communities for all.

4. HowGood

Founder: Alexander Gillett

“We can solve the climate crisis largely via the food system. Because not only does it produce the most [carbon emissions], but it is one of our best opportunities for sequestering carbon. [At HowGood], we map it out, we make it easy. We take the data workload off the hands of the people who want to be implementing the change.”

Under the leadership of Alexander Gillett, co-founder and CEO, HowGood is revolutionizing sustainability intelligence for the food industry. With the world’s largest product sustainability database, HowGood enables companies to make informed decisions that reduce their carbon footprints and promotes environmental stewardship. Learn how HowGood’s Latis platform offers granular insights that drive sustainability reporting and carbon reduction efforts, allowing companies to easily implement change.

Hear from Alexander Gillett, co-founder and CEO of HowGood, who talks about offering the world’s largest product sustainability database at your fingertips.

As we conclude our Earth Month showcase, let’s continue to support and celebrate startups like Clarasight, Revel, BlocPower, and HowGood that are driving positive change for our planet. Together, we can build a more sustainable and resilient future for generations to come.

AWS Activate credits now accepted for third-party models on Amazon Bedrock

Howard Wright — Tue, 02 Apr 2024 11:48:43 +0000

If there’s one thing we’ve learned at AWS, it’s the importance of experimentation. When you’re creating something new, it’s crucial to be able to try out different technologies and quickly iterate on an idea. But experimentation can be expensive, especially for a scrappy startup team at the earliest stages of innovating. That’s one reason why we launched AWS Activate, a program focused on supporting startup founders in every step of their journey, including providing more than $6 billion in credits to help you and others like you experiment on the AWS cloud with little-to-no upfront cost.

Today, we’re taking another step to make it even easier for founders to build and iterate on their solutions using the latest technologies. We’re making AWS Activate credits redeemable for third-party models on Amazon Bedrock, our fully-managed service that offers a choice of high-performing foundation models (FMs) from leading artificial intelligence (AI) companies, like AI21 Labs, Anthropic, Cohere, Meta, Mistral AI, Stability AI, and Amazon via a single API. This means founders everywhere can now use their AWS Activate credits to experiment with these and other FMs, along with a broad set of capabilities needed to build responsible generative AI applications with security and privacy. Our goal is to make it easier for startups to evaluate what FMs are more appropriate for their use cases and find the perfect match.

Helping startups innovate with generative AI

With our full-stack generative AI offering, AWS is helping more companies around the world embrace the potential of generative AI to transform customer experiences, enhance people’s productivity, and discover new business opportunities. Expanding AWS Activate credits to Amazon Bedrock is a key way to help startups leverage generative AI in their solutions from the start.

This is one of the many benefits we have identified by working backwards from the needs of our customers, like Y Combinator. Since 2005, Y Combinator has funded and helped founders launch, build, and scale over 3,000 companies—including 60 unicorns—which currently have a combined valuation of more than $600 billion.

“With virtually every startup quickly becoming an AI startup, our partnership with AWS has never been more relevant to the companies getting into our program,” said Michael Seibel, Group Partner at Y Combinator. “AWS has been a long-standing partner and a relentless advocate for our founders, helping them with hands-on support and access to the tools they need to build the products and services people all over the world use and love.”

Since 2009, AWS has provided hundreds of millions of dollars and technical support to Y Combinator companies such as Stripe, Brex, Rappi, and many more—and in the last three years alone, AWS has provided more than $125 million in credits to Y Combinator startups. Our high-touch approach has resulted in Y Combinator companies consistently choosing AWS as their cloud provider—more than 70% of Y Combinator-funded companies run on AWS and if you look at the last two years alone, when the use of AI/ML has become more prevalent among startups, that number jumps to 80%.¹

For the latest Y Combinator cohort (January 2024), AWS put together an exclusive package of benefits to help startups reduce upfront costs and get access to reliable, high-performance infrastructure to build their generative AI applications on. This includes $500,000 in AWS credits that can be used for:

AWS Trainium, our purpose-built chip for training deep learning models, which offers up to a 50% cost-to-train savings over comparable Amazon Elastic Compute Cloud (Amazon EC2) instances;
AWS Inferentia, a chip designed to enable models to generate inferences more quickly and at lower cost, with up to 40% better price performance;
Reserved capacity of up to 512 NVIDIA H100 GPUs via Amazon EC2 through Capacity Blocks for Machine Learning, which dramatically increases GPU availability and ensures startups have reliable, predictable, and uninterrupted access to the GPU compute capacity required for their critical machine learning (ML) projects;
and now third-party FMs on Amazon Bedrock.

Foundation models for all

Any startup can join AWS Activate and apply for up to $100,000 in AWS credits. In addition to experimenting with Amazon Bedrock, AWS Activate credits can be used to offset costs of AWS services, including infrastructure technologies like compute, storage, databases, AI/ML, and more. Learn more and become an AWS Activate member at startups.aws.

¹Numbers as of February 2024.

Unlocking the value of unstructured data: How Coactive built a visual analytics platform on AWS

Megan Crowley — Mon, 01 Apr 2024 15:15:52 +0000

It is said that a picture is worth a thousand words – and, according to Forrester Research, a minute of video may be worth 1.8 million words. For businesses ranging from ecommerce to social media, visual content is worth more than the amount of words that it conveys: it is an opportunity to build customer engagement, increase trust and safety, enhance personalization, and glean actionable insights based on content engagement.

Coactive, a visual data analytics startup founded by CEO Cody Coleman and Will Gaviria Rojas in 2021, is democratizing the opportunity for businesses to analyze images and videos.

Coactive’s co-founders Will Gaviria Rojas and Cody Coleman

Images and videos are unstructured data—information that has not yet been ordered in a predefined way—that traditionally require machine learning expertise, a robust technical infrastructure, and significant amounts of time to accurately analyze.

Coactive, the platform built on Amazon Web Services (AWS) and available on the AWS Marketplace, helps data practitioners derive rapid insights from unstructured data at scale and with minimal supervision. Accessible by user interface or APIs, the platform’s capabilities range from intelligent search to production analytics that use the full power of SQL.

Proving what’s possible in AI

Coactive’s innovative solution results from intensive amounts of time, research, and determination. During 2018, while earning his PhD in Computer Science at Stanford, Cody recognized that “artificial intelligence and intelligent applications were going to be the future. The blocker was that you needed hundreds of thousands of dollars’ worth of equipment and tremendous amounts of data to accomplish anything significant.”

Use SQL to run analytics on your visual dataset. This photo shows a SQL query used to categorize images

Bothered by these limitations, Cody committed to lowering the barriers to entry for machine learning so that everyone could benefit from it: “My mission during graduate school was to use my passion for computer science to benefit society at large while serving as a leader for future generations.”

Cody joined the Stanford DAWN research project, a group focused on making it dramatically easier to build AI-powered applications. One of the many impressive breakthroughs from Cody’s work was DAWNBench, the first end-to-end machine learning (ML) systems benchmark used by global technology companies as the industry standard. In its first year, DAWNBench reduced model training time by 500x and training cost by 20x. Galvanized by the progress he’d made in creating accessible AI, Cody tackled the next big question: What to do next?

At this time–by serendipity or coincidence–Cody’s friend Will moved to the San Francisco Bay area to begin a career at a large technology company. With a friendship spanning 10 years since their time as undergrads at Massachusetts Institute of Technology (MIT), Cody helped Will move in. “Will asked me the two questions you should never ask a PhD student because they cause an immediate existential crisis,” laughs Cody. “’When are you going to graduate?’ and ‘What are you going to do after?’”

Cody had considered options that ranged from joining a prestigious technology company, becoming a university faculty member, joining a startup, or building a company. “Without hesitation, Will told me to build my own company,” says Cody. “He told me that it’s the right time and I have the right knowledge. And that he’d love to join me in this journey.”

Months of conversations, research, and studying the problem first-hand led Cody and Will to come to the same realization: People need a visual analytics platform to unlock the value of their content, and it’s time to build it. With that decision, Coactive was founded.

The Coactive team works together on version 1.0 of the Coactive platform.

Creating a visual analytics solution for everyone

Machine learning advanced significantly during Cody’s time at Stanford, but there was still much work to be done to make AI applications accessible to everyone: from the world’s largest companies, to a startup designing their minimum viable product.

This was particularly true about machine learning that analyzed images and videos to derive actionable insights. For these unstructured data formats, the end-to-end workflow could require high-end large-scale compute in the form of GPUs, significant storage capacity, and large amounts of time and expertise due to the complexity of the process. A common workflow may include the following:

Data scientists complete data exploration and build computer vision models to analyze and understand the visual data.
ML engineers operationalize these models.
Software engineers plug the model predictions into real world applications for consumers.

To make visual content analysis more accessible, accurate, and efficient, Coactive pairs the breadth of existing large language models (LLMs) with the accuracy and automation that comes from applying a learning system to domain-specific data. After customers provide access to their large volumes of raw image and video files, Coactive uses pre-trained foundation models in conjunction with their proprietary active learning and classification system to embed and index the data. During this process, customers have the option to upload existing labels or provide a few examples so the Coactive platform can further learn any domain-specific nuances of their data.

“One of the very powerful things about really large models is that we don’t actually need to toss a massive quantity of data to fine-tine for specific tasks,” explains Cody. “They call these large language models ‘few-shot learners’ for a reason. Rather than thinking about the quantity of data we toss at these systems, it’s really about quality.”

The result? Customers can use Coactive to query, search, filter, and analyze visual content rapidly and at massive scale.

Partnering with AWS to accelerate success

As an innovative and rapidly-scaling startup, Coactive decided to migrate from their original cloud provider go all-in on AWS. Solutions offered by AWS align with Coactive’s four primary cloud provider needs: Depth and breadth of services, optionality in tooling, availability to power the scaling of their product, and security-first offerings.

“We needed to build our solution on a cloud provider that could handle enterprise scale while being flexible enough to let us create something entirely new. With AWS, we were able to do this while ensuring best-in-class security to our customers,” says Cody.

Building with AWS solutions

After migration, Coactive set to work building a cutting-edge web application using AWS solutions such as Amazon Simple Storage Service (Amazon S3), Amazon Aurora, and Amazon Elastic Container Service (Amazon ECS). This web application helped Coactive establish their initial MVP and run proof of concepts for prospects.

For their data-centric machine learning jobs, Coactive benefits from using Amazon Aurora PostgreSQL Serverless to serve low latency database requests without having to spend time managing their database infrastructure. Coactive’s many petabytes of image and video data are stored using Amazon S3.

To front their web application, Coactive uses a combination of Amazon CloudFront as its content delivery network (CDN). The backend web application runs on Amazon ECS, communicating with their database and peripheral downstream services such as Databricks on AWS. Amazon ECS provides Coactive simplicity in managing the container infrastructure running on Amazon Elastic Compute Cloud (Amazon EC2).

Security and data privacy are critical aspects of machine learning workloads. To provide their customers with a secure analytics experience, Coactive uses Amazon GuardDuty, Amazon Inspector, AWS Key Management Service, and more. With these solutions, Coactive achieved SOC2 cybersecurity compliance over the course of a single quarter.

Bringing their product to market

It is important to ensure a successful go-to-market motion. To share their product with the global audience of AWS customers, Coactive joined the AWS Partner Network (APN) and lists their product on the AWS Marketplace.

Coactive is also a member of the AWS Global Startup Program (GSP), offered through the APN. This program pairs Coactive with an AWS Partner Development Manager who provides support in three key areas: product development, go-to-market, and co-selling.

Accelerating success with AWS Startups

In addition to building with the help of AWS technical solutions and business support, Coactive leverages the AWS Activate program. AWS Activate provides startups with resources ranging from credits and exclusive offers, to technical support and networking events.

In collaboration with the AWS Startups team, Cody and other AWS Activate members recently shared their expertise at AWS GenAI Day, a one-day virtual event showcasing how startups are building with generative AI on AWS. As part of the keynote panel “Mapping the Trajectory of GenAI: From Learning to Impact,” Cody explained why data is a critical piece of generative AI and how recent breakthroughs in machine learning have the potential to significantly improve lives.

Panelists discuss the impact of generative AI during the “Mapping the Trajectory of GenAI: From Learning to Impact” session.

Building for the future

Coactive continues to build a product that lowers the barrier to entry for machine learning and Cody notes that proving what’s possible—and helping other people to prove it as well—is an important part of his mission. His incredible story includes being born during his mom’s incarceration, placed into foster care, and adopted by grandparents who raised him within the constraints of economic inequality. As the first Black PhD student to graduate from Stanford in nearly 20 years, Cody is familiar with the challenges of being an underrepresented person in technology. He’s committed to making diversity, equity, and inclusion core principles at Coactive. “How we succeed is just as important as the fact that we do succeed,” says Cody.

“Will has a great saying where he says his goal in life is to make ladders so it’s easier for people to follow in his footsteps,” Cody explains. “My mission in life in general is to demonstrate that regardless of where you come from, you can be successful. If I could do it, anyone can do it.”

For people who are considering founding a startup, Cody shares that fear is normal: that you’re not cut out for the CEO role, that it’s a big risk to start a company, that things will be hard for a long time. His moment of confidence came when he realized, “I don’t need to have everything figured out to get started. I just need to start to figure everything out.”

Two years later, the success of this advice is evident. The Coactive team continues to level the AI playing field by bringing impactful visual analytics to their customers. Cody’s commitment to making data useful remains strong. “One of the most amazing use cases I’ve seen is fine-tuning a speech recognition model to recognize the signs of a dangerous respiratory condition in a baby’s cry,” he explains. Early detection using AI reduced the infant mortality rate and lowered the time, cost, and skill necessary to make an accurate diagnosis.

“Sustainable and ethical AI has incredible potential to meaningfully improve lives,” says Cody. “One of my biggest motivators to this day is that by democratizing AI with companies like Coactive and AWS, there are so many stories people are going to tell and questions they are going to answer. I’m excited to see it.”

How Flodesk Is designing a new narrative for startups

Bonnie McClure — Mon, 25 Mar 2024 19:25:24 +0000

For the creators of Flodesk, your inbox is personal.

The San Francisco-based startup, founded in early 2018, is designing emails people love to get. As co-founder Rebecca Shostak says, “Flodesk is a relationship-building software. I want everybody to harness the power of these one-on-one, intimate relationships that they can have with their followers.”

Rebecca’s design background is rooted in rock n’ roll; she began her career designing merchandise for a company that needed, in her words, “a female touch” for the artists on their roster. It was her first exposure to designing for different types of brands—and creating beautiful designs that sold in a variety of styles. These are the skills that Rebecca brought with her on her first solo venture: a template shop designed for popular email platform MailChimp.

“I had always wanted my own business. I’m from Silicon Valley, and I have entrepreneurship in my blood from my dad. I wanted to start my own company. I saw a niche for Photoshop templates for wedding photographers. And there was a huge demand for it,” Rebecca explains. When she ran into customers struggling to implement her designs into the MailChimp platform, a bigger idea was formed: a platform that allows people to create beautiful, branded emails easily. “All email marketing companies were eco-oriented, focused on workflows and integrations. They weren’t focusing on the brand and the design,” Rebecca says. “I was like, ‘I have to do this.’ It was almost a religious calling.”

Meanwhile, Flodesk’s co-founder CEO Martha Bitar was working in customer relationship management (CRM) for small business owners, in a similar market. “Martha was marinating in the same customer ecosystem that I was, just in different lanes,” Rebecca explains. Both Martha and Rebecca sensed a common problem in their world: “We see these people who are making millions of dollars with their services and their website looks amazing, their Instagram looks amazing, their photos look amazing, and their emails suck.”

Designing a new narrative for startups

For Flodesk, it has always been about flipping the narrative. The company has strong roots in design—they were an early leader in dreaming big about how beautiful emails can be. The oxymoron formed by “beautiful” and “emails” built the foundation of what would become Flodesk, according to Rebecca: “We’re taking softer products that, in the past, have been considered clunky, ugly, difficult to use, and just unappealing. For the small business market who’s extremely brand-focused, we’re flipping that on its head and making something that people never thought could be sexy, really sexy.”

But the company’s disruption doesn’t stop at design. When Martha and Rebecca came together to create their prototype around January 2018, they turned to the path relatively untrodden for startups. As Rebecca details: “we thought, what if we pitched this and did it without funding? What if we were able to validate it and get the software going without even taking venture capital money?” Leaning into this concept, by the summer of 2019, Flodesk had a prototype. When an influencer friend of Rebecca’s built and sent an email using Flodesk, adding a plug for it at the bottom, trial requests began flooding in. “Flodesk launched itself long before we had the actual launch,” says Rebecca.

Building a trusted brand through AWS

Flodesk turned to Amazon Web Services (AWS) as a partner from the very beginning—the platform was built on Amazon Simple Email Service (SES). “I can’t imagine Flodesk without AWS intertwined with it,” says Rebecca. “They’re like the giant whose shoulders we’re standing on, in some sense, for our tech.”

In the early days, the founders faced infrastructure problems with a sudden influx of subscribers. As Rebecca details: “we started out with AWS and we built on them. We were so open at the beginning with our policies because we just wanted as many people in the door as possible, but we found that that let a lot of scammers in. We were so brand new, things were going so quickly—we weren’t able to put the proper barriers in place to get the scammers out at first.” Rebecca spent a lot of late nights on the phone with AWS to troubleshoot the issue. “Early on, there were a lot of really close calls,” she explains. “We had moments where we were getting our server shut down because of scamming. But since we’ve worked with AWS and partnered with them to advise us on how to build out a more and more robust trust and safety team, that’s led us to be where we are today. I just can’t imagine our story without AWS in it.”

Tony Silva, Flodesk’s startup portfolio lead at AWS, was there to answer those late-night calls from Rebecca. Tony now partners with Flodesk’s engineering team internationally to troubleshoot any issues that may pop up. As Tony explains, “with Flodesk, our relationship is outside even the tech perspective at this point. It’s about their entire infrastructure. It’s helping them lower their costs, making sure that they’re architected the right way, providing any sort of support, whether it’s a go-to-market or a personal relationship or a touchpoint.”

For small businesses, the future is bright

Today, with 70,000 customers, over $20 million in annual recurring revenue, and 35 team members all over the world, Rebecca is still adjusting to enormous growth: “sometimes I still feel like I have no idea what I’m doing. But I think we always had that bootstrap mindset.”

Rebecca attributes a great deal of Flodesk’s success to the company’s founding principles. A focus on problem-solving for their customers continues to guide the company today. As Rebecca says: “we started with a founder idea. We had a product-market fit before we even touched a line of code. Start with the problem, not the solution. We were really focused on this segment of the market that’s been left behind by the legacy players. We wanted to create an experience where you could self-serve, that you weren’t reliant on going back and forth with our team. And one that was affordable too.”

For Rebecca, the future is “wildly exciting” for small businesses. “I really believe that small businesses, and the creative people that run them, are going to own the future. And that future really excites me,” she says. Her outlook for Flodesk is particularly bright: “I want Flodesk to be the household name that they associate with growing business. My vision is that you have a household, you have an Amazon Prime account, you have a small business, and you have a Flodesk account.”

Rebecca’s advice for startup founders? “Be scrappy. Sometimes the best ideas come from the scrappiest of places. Use your imagination and be creative.”

AWS and IRCAI announce winners of the 2023 Compute For Climate Fellowship and open applications for 2024

Lisbeth Kaufman — Thu, 21 Mar 2024 12:07:44 +0000

Applications are open through June 7, 2024 for the first-of-its-kind program to fund proof of concepts for new ideas that leverage advanced cloud computing, including generative AI and high performance computing, to solve some of the biggest challenges in the fight against climate change.

How can the world speed up efforts to address the climate crisis? Climate change is already threatening lives and disrupting businesses through increased wildfire, extreme heat, flooding, sea-level rise, and drought. With the news that 2023 was the hottest year on record, the race is on to get to net-zero-greenhouse gas emissions as soon as possible.

Speed is critical to addressing the climate crisis before even more damage is done. Luckily, Climate Tech startups are inventing brilliant new technologies and business models to address the climate crisis. At AWS, we want to help Climate Tech startups move even faster with our advanced computing services, like high performance computing (HPC), artificial intelligence/machine learning (AI/ML), and generative AI. That’s why in 2023 we launched the Compute for Climate Fellowship in partnership with the International Research Centre on Artificial Intelligence (IRCAI), an organization under the auspices of the United Nations Educational, Scientific and Cultural Organization (UNESCO).

The Compute for Climate Fellowship is a first-of-its-kind global program to fund groundbreaking proof of concepts (POCs) that think big, innovatively use advanced cloud computing, and enable the world to more quickly address the climate crisis. We are thrilled to announce the startups accepted into the Compute for Climate Fellowship in 2023 and to share that we’ve opened applications for the 2024 fellowship, which can be submitted here.

The 2023 selection process was very competitive. We received applications from 24 countries and choose four Fellows, a 5.7% acceptance rate. Our finalists include startups working on solutions that feature fusion energy, ocean carbon measurement reporting and verification (MRV), water purifying chemicals, and drought resilient crops. These four startups are building truly innovative projects that have the potential to significantly advance the fight against climate change.

Coastal Carbon

The ocean is one of the world’s largest carbon sinks. Measuring sequestered ocean carbon can be the hardest and most expensive part of an ocean carbon project. Coastal Carbon is building AI models to solve that. The startup’s models can see underwater from satellites to track and measure underwater vegetation growth and how much carbon the vegetation captures. Their AWS-funded POC will use Amazon SageMaker and AWS HPC services to train satellite-agnostic foundation models of earth and sea that can identify and quantify ocean blue carbon. Their training corpus includes more than 50 years of geospatial data, which enables their AI models to generalize with greater accuracy. The Compute for Climate Fellowship POC will help Coastal Carbon increase transparency and trust in ocean carbon removal. They estimate that it will also facilitate up to a 1,000x increase in ocean monitoring and data collection capacity, which was previously done manually by divers and sensors.

“We have accelerated our product development thanks to this program, which has allowed us to make even bigger plans for the scale of Coastal Carbon,” said Kelly Zheng, CEO of Coastal Carbon. “Remediating climate change is urgent. A faster POC means more tonnes of carbon removed that is measurable, scalable, and affordable, with the highest scientific rigor.”

Phytoform

With drought, flooding, and extreme heat disrupting agriculture worldwide, the world needs crops that can withstand the impacts of climate change. That’s where Phytoform comes in. Phytoform is operating at the edge of biotechnology, AI, and agriculture, developing novel crop genetics to create more climate resilient plants. As a winner of the Compute for Climate Fellowship, Phytoform will use the program to enhance the accuracy of CRE.AI.TIVE, their unique AI-powered pipeline that can draw novel insights from plant genomics, evolve new crop characteristics, and make plants more resilient. The POC will help Phytoform find new research and development (R&D) solutions for more climate resilient seeds at scale.

“The Compute for Climate Fellowship is absolutely crucial for Phytoform’s mission of improving sustainability in agriculture,” said Nicolas Kral, CTO of Phytoform. “With the help from AWS and IRCAI we were able to really accelerate our solution to deploy a unique approach to enhance R&D and secure the food of tomorrow.”

Realta Fusion

Realta Fusion is developing compact magnetic mirror fusion energy systems to decarbonize industrial process heat and power. Their aim is to create the fastest path to commercial deployment of fusion energy. Through the Compute for Climate Fellowship, Realta Fusion is developing the first-of-a-kind plasma stability simulation in the cloud by leveraging Amazon Elastic Compute Cloud (Amazon EC2) HPC instances and Elastic Fabric Adapter. Plasma stability is a principal design requirement for fusion power plants. This is groundbreaking. Prior to this POC, there were only two super computers in the United States that could handle these simulations, and they are at National Labs with a year-long waitlist. With this POC, Realta Fusion will demonstrate that the plasma stability simulations can be run in the cloud, which will help democratize access for the whole fusion industry.

“Advances in AWS cloud computing have paved a new pathway for Realta Fusion to expedite research in magnetic confinement fusion energy systems.” said Cary Forest, PhD, co-founder and chief scientific officer at Realta Fusion. “Such intensive plasma physics simulations have never been done in the cloud before. The Realta Fusion team is grateful for this fellowship award and excited to use the power of cloud computing to advance fusion energy research.”

Xatoms

Xatoms is a water treatment startup using AI and quantum computing to discover molecules of substances that can purify polluted water. Their Compute for Climate Fellowship POC is focused on creating a water purification technology that is both effective and affordable, aimed at assisting those who currently lack access to safe drinking water. Xatoms’s proposed POC will target end users in developing countries, particularly underserved communities lacking clean water access, aiming to reduce mortality rates and redirecting time spent fetching water towards education and entrepreneurship. Their POC will develop a new AI/ML algorithm capable of analyzing a large amount of chemical data, followed by simulations of novel molecules that can purify water. They will use AWS services like Amazon Braket for storage and quantum simulation software and Tangelo to accelerate the discovery process.

“The Compute for Climate Fellowship will significantly boost our material discovery efforts, thanks to the resources and mentorship provided,” said Diana Virgovicova, CEO of Xatoms. “We are proud to be backed by some of the biggest names in the industry, highlighting the urgency of addressing global water challenges affecting more than two billion people worldwide.”

Applications for 2024 Compute for Climate Fellowship are now open!

We are thrilled to share that startups can now apply for the 2024 Compute for Climate Fellowship here. Applications will be accepted until 11:59 pm June 7, 2024. Applications submitted after that date will be considered for the 2025 Compute for Climate Fellowship.

We are making a few changes to the fellowship this year to make it a resource for even more Climate Tech startups. Like last year, the program is global and startups from all countries around the world are invited to apply. This year, we are expanding the scope of proposals to address at least one of these eight solution areas:

Clean energy
Low-carbon transportation
Sustainable agriculture and food
Circular economy/manufacturing/industry
Sustainable buildings
Greenhouse gas accounting and sustainability management
Carbon removal
Environment (water, pollution, biodiversity) and climate risk

This year, we are looking for projects of various sizes and sophistication. AWS will fund up to $1.5M in total AWS credits for the Compute for Climate Fellowship. We will look for proposals that can be completed in 2-3 months and that think big, creating novel solutions to address the climate crisis.

Through the fellowship, IRCAI and AWS will provide global climate tech startups access to various technical resources to build their POCs, including:

A team of AI, sustainability, and ethics experts
Access to advanced computing services, such as HPC and quantum computing
Cloud computing products and services that support AI, generative AI, and ML solutions
AWS Activate credits to cover the POC build

In addition, all POCs will be designed under the guidelines of UNESCO’s Ethics Impact Assessment for AI. This ensures that each solution is built with safe, trustworthy technology.

Startups that apply but are not selected to participate will have access to up to $5,000 in AWS credits. They will also be invited to join the IRCAI industrial club.

To apply for the Compute for Climate Fellowship and find out if your startup is eligible, visit https://ircai.org/compute-for-climate-fellowship.

How Tiffany Johnson is working to democratize access and define success for underrepresented founders

Bonnie McClure — Wed, 28 Feb 2024 14:37:35 +0000

To celebrate Black History Month, AWS Startups is excited to spotlight Tiffany Johnson, global business development manager on the AWS Startups underrepresented founders business development team, who has spent her eight years at Amazon challenging customer perceptions and building bigger and better.

When Tiffany Johnson started on Amazon’s sales team in 2019, she identified a large gap in the needs of the customers that she was working with. It was after a year and a half of customer research that she approached her VP with two of her colleagues, Rachad Lewis and Jeremy Erdman, to deliver the findings: “we took the time to understand how Amazon was supporting Black-owned businesses, what some of the gaps these businesses are facing, and determining whether we had the solution to help them with those specific needs.” The team’s research would become the Black Business Accelerator, a $150 million commitment by Amazon that empowers sustainable entrepreneurship for Black-owned businesses.

Tiffany is the first to admit that the road to developing such a herculean commitment was anything but easy, but credits her team at Amazon for their success: “I am of the belief that if you want to go fast, go alone, but if you want to go far, go together. And I could not bring this project to life without the support of a hundred-and-something people who believed in what we were building.”

Today, as a business development manager on the underrepresented founders team at AWS Startups supporting Latino, Black, Native American, and women founders, Tiffany is still focused on understanding the challenges that underrepresented startups face. As a part of AWS Startups, Tiffany’s team engages with underrepresented founders and investors through partnerships, events, programs and a variety of high-touch engagements, looking to create tools and resources to democratize access to some of those hiccups that they pinpoint.

Bridging the funding gap

One such hiccup is access to funding, an inequity that is striking in the world of startups. In 2022, women-only led companies received only 1.9% of venture capital funding, while Black and Latino-founded companies received only 1%. One founder working with Tiffany’s group joined a pitch competition only to realize he was the only entrant that was lacking in a special type of first round funding: the friends and family round.

“When we think about the systemic challenges that underrepresented founders often face, sometimes it’s the little things that people don’t think about,” Tiffany says.

After identifying this obstacle, the underrepresented founder team quickly connected him to investors in their network. The team also partners with different organizations that host these pitch competitions, bringing in AWS credits to help underrepresented founders offset costs while they’re building on AWS.

Building social equity through mentorship

Another hurdle seen firsthand by Tiffany is guidance from experienced entrepreneurs and business leaders. To bridge that gap, her team has created mentorship dinners that partner with different C-Suite executives and leaders in the ecosystem to connect with founders.

“I think you’re only as strong as your network, and it starts with mentorship,” Tiffany says. “It’s always good to have that person who has experienced the path that you are treading, to be there as a guide. They come with that relevant experience that you can resonate with and that can help get you to the next step.”

It’s a tenet that Tiffany credits with her own journey as a Black female business owner: “I’m the first person in my family to ever work in corporate America. Navigating corporate America can be extremely challenging, but to have someone who has that wisdom to guide you who can be a sounding board for the good and the bad has been helpful in reaching that next step in my career, or even just closing out some of the projects that I’ve been able to launch internally at Amazon.”

AWS Reach is one such project from Tiffany’s group that puts mentorship for underrepresented founders to practice. AWS Reach is a global community that provides underrepresented founders access to $5,000 in AWS credits, marketing support, and events throughout the year that meet their specific needs.

Leveraging AI/ML for underrepresented founders

When it comes to the future for underrepresented founders, Tiffany is optimistic despite a shifting venture capital landscape that could disproportionately affect the startups that she works to help. One trend she’s most excited about: artificial intelligence and machine learning (AI/ML).

“This is a fast-moving field, and we don’t want underrepresented founders to get left behind,” Tiffany says. “There are so many opportunities for underrepresented founders to make their mark in this space.”

For Tiffany’s team, that means working with founders to see how they can help them eliminate biases from AI/ML models, and helping them innovate in some of the markets that are available that other players aren’t thinking about. In leveraging some of the technologies that AWS already has to offer, like Amazon Bedrock, a fully managed service that offers access to high-performing foundation models from leading AI startups and AWS, the underrepresented founders team is scanning the landscape to identify where founders can get a leg up in new technologies.

Be fearless

When asked what advice she has for Black founders and entrepreneurs building on AWS, Tiffany emphasizes one thing above all: don’t be afraid to seek out the opportunities. “There are so many opportunities internally here at AWS for underrepresented founders. So, don’t be afraid to seek them out and fully familiarize yourself with them.” That means leveraging all the resources AWS’s programs have to offer: “Founders come in and they cherry pick—take this, take that. Leverage all of the resources and stay engaged, stay consistent, and you’ll reap the benefits of what AWS is offering to support underrepresented founders.”

How climate tech startups use generative AI to address the climate crisis

Lisbeth Kaufman — Mon, 19 Feb 2024 21:41:05 +0000

Generative artificial intelligence (AI) can seem like magic or like a clever collaborator, generating original text, images, videos, and music. It has grabbed the world’s attention with incredible chat capabilities and captivating image creation. However, it can be more than a creative collaborator or a chat bot. At AWS, we are seeing generative AI transform how humans and business use technology to solve some of the world’s most challenging problems.

There are few problems more urgent than the climate crisis. The world is in a race to get to net zero carbon emissions by 2050 to curb global warming to 2 degrees Celsius before the impact of climate change becomes irreversible. Speed is critical to addressing the climate crisis— generative AI is still an emerging field, but it is already becoming an important tool to accelerate the build and deployment of climate solutions.

We are excited to introduce you to a few Climate Tech startups that are at the forefront of the race to stop the climate crisis. They’re using generative AI to fight climate change by reducing greenhouse gas emissions and enabling the world to transition to a zero-carbon economy.

BrainBox AI: Accelerating building decarbonization through generative AI

According to the International Energy Agency (IEA), buildings account for 30 percent of global energy consumption and 26 percent of global energy-related emissions. Reducing buildings’ energy use is critical for getting to net zero global emissions.

BrainBox AI has developed autonomous AI to decarbonize and optimize commercial buildings. It also saves customers money on their energy bills. The cloud-based optimization solution built on AWS connects to buildings’ existing HVAC (heating, ventilation, and air conditioning) systems and autonomously sends real time optimized control commands to minimize emissions and energy consumption without any human intervention.

For example, BrainBox AI has helped building owners reduce HVAC energy costs by up to 25 percent and reduce HVAC-related greenhouse gas emissions by up to 40 percent by predicting the temperature in a retail store based on historical data and external datasets, like weather and energy tariff structures.

As BrainBox AI adds new buildings to their system, they utilize generative AI to reduce onboarding time for each new building. In the past, whenever a new piece of equipment was identified in a building, such as a pump or an air handling unit, engineers had to go through the complex technical manuals from the manufacturer, find details like the pump’s power rating or the pressure it generates, and finally convert that information into a machine-readable format.

Using Amazon Bedrock, BrainBox AI extracts data and generates configuration files automatically. These files are then completed and revised by engineers. This process is known as power tagging. With Amazon Bedrock, BrainBox AI estimates they reduced the time it takes to power tag by over 90 percent. As a result, BrainBox AI is able to onboard more customers, more quickly, so it can have a bigger, faster impact on the climate crisis.

BrainBox AI shows energy optimization of a building

Pendulum: Decarbonizing the supply chain with generative AI

Pendulum harnesses the power of AI to address one of the world’s most pressing problems: how organizations can create more from less. The company’s technology offers sustainable solutions for complex problems in sectors such as commercial supply chain, global health, and national security.

Optimizing supply chains is crucial to reducing carbon emissions. Accenture estimates that supply chains generate 60 percent of all carbon emissions globally. According to the U.S. Environmental Protection Agency, supply chains can account for more than 90 percent of a company’s greenhouse gas emissions. When you look at how supply chains work (or don’t work) today, there are some startling details about the extent of wasted resources and capital. For instance, every year, an estimated $562 billion is lost in overstock, with 17 percent of food products and 8 percent of retail and consumer packaged goods products being discarded.

Pendulum’s AI-powered solutions enable organizations to intelligently manage their operations and reduce product waste, revenue loss, and excess greenhouse gas emissions. Built on AWS, Pendulum’s software can predict demand, plan supply, and geolocate shipments. It allows companies to more accurately purchase the resources they need while producing exactly the amount of goods their customers demand.

Accessing enterprise data is critical for Pendulum’s platform. However, data is often retained in siloed systems and unstructured documents such as PDF and plain text files. Pendulum’s software is designed to leverage the data sources most relevant to operational decision-making. They are deploying generative AI to rapidly unlock important information contained in long and complicated documents so they can accelerate time to value for their customers.

One example of where this is being effectively deployed is in precision agriculture. The Pendulum team uses a human-in-the-loop approach to instruction-tune a large language model (LLM) on AWS Trainium using Amazon SageMaker. This generates machine-readable data from unstructured files that their customers’ farming machinery can use to determine how much pesticide, water, and other products to use. As a result, the customer is less likely to overuse or over-order resources, can save money, and reduce their carbon footprint and environmental impact. Furthermore, it allows customers to adhere to plant needs, local regulations, and other important criteria.

Pendulum estimates this solution has reduced the time required to decode these documents by 83 percent, and they now only need to review the data for quality assurance. This in turn reduces costs and accelerates the deployment of their software at scale.

Precision agriculture

VIA: Making it easier for building managers to understand energy efficiency with generative AI

To enable emissions reduction, institutions and businesses need to track energy data at the local and individual level. For instance, in order to reduce the carbon emissions associated with its fleet of electric vehicles (EVs), it’s important for a company to understand if EVs, in a certain region, at a certain time, are charged using electricity powered by renewables or fossil fuels. For energy efficient buildings, in-depth individualized data across an organization’s entire real estate portfolio is essential. If everyone provided all data transparently, this wouldn’t be a problem. However, individual-level data is often not accessible because of privacy issues or security concerns. Many individuals are reluctant to provide the time, date, and location of their vehicle charging/discharging/energy data. This makes energy management and greenhouse gas reduction challenging.

Via Science, Inc. (VIA) enables organizations to reduce their carbon footprint as a collective, while keeping individual data private and secure. The company provides sustainability data using zero-knowledge proofs tested and verified by the U.S. Department of Energy. This enables organizations and businesses to track data and meet sustainability goals even when it’s not possible to share detailed information due to regulatory or privacy barriers.

VIA initially developed a solution for the U.S. Air Force, which has strict data privacy requirements that often prohibits building management and energy management teams from accessing critical data they need. VIA’s decentralized software solution enables airmen and permitted contractors to use generative AI models without sharing data: no private data is used to train the model or sent to the model in the prompt. Instead, when a user enters a prompt like “show me all buildings on Air Force Base XYZ with HVAC system condition less than 60,”, the LLM responds with “I understand what you want to achieve, and, because I don’t have access to the data, I will generate a SQL query that you can run to get the data from your local database. I will also send you the frontend code you can run to display the data.” These two pieces of code are then sent back to the user where the tool, SLAM AI, automatically runs and visualizes the data locally.

To further save energy and reduce compute costs, VIA uses compact open-source LLMs that run on CPUs. They continually assess new models due to the rapid evolution of LLM performance. Leveraging Amazon Elastic Kubernetes Service (EKS), they can seamlessly hot swap models, integrating more efficient ones as they become available.

Interface of VIA’s SLAM AI tool

What’s next for generative AI and climate tech

BrainBox AI, Pendulum, and VIA are using generative AI on AWS in exciting ways to address the climate crisis. They make use of generative AI’s ability to extract key elements from unstructured data and generate new content. This enables these companies to serve their customers more quickly, serve more customers, and reduce greenhouse gas emissions. It also reduces costs for these companies and for their customers.

We expect that Climate Tech startups will find additional new ways to use generative AI on AWS to address the climate crisis. Here are a few examples of what we are seeing in other industries that we think could apply to Climate Tech.

Data augmentation using generative AI to generate synthetic data for predictive model training

Generative Al can create synthetic data, which is a class of data that is generated rather than obtained from direct observations of the real world. This could be useful for subsurface modeling for geothermal or carbon sequestration where subsurface rock formation data is hard to come by. Startups in low-carbon transportation could also use generative AI to create scenarios to test new vehicles. It could also be useful in Climate Tech hard-tech manufacturing. Synthetic image data creation can be used for creating images of equipment (e.g., compressors, turbines) with rust or cracks. These images can be used for training vision-based machine learning (ML) models for predictive maintenance, which can play a key role in reducing costs and minimizing operational downtime.

Improve Climate Tech manufacturing efficiency using generative AI

By using models trained on historical data, including machine usage and maintenance logs, generative AI can identify patterns and links between various factors, such as temperature, vibration, and operating hours. This can enable the system to foresee the likelihood of equipment failure and proactively communicate those patterns to the right stakeholders such as quality engineers, maintenance engineers, and operators. By proactively communicating the need for maintenance, downtime will be reduced, minimizing disruptions to manufacturing.

Design and synthesize new protein sequences for sustainable agriculture and food production with generative AI

Generative AI can predict protein folded structures that enable them to carry out particular functions in the cell. This will allow researchers to generate functional proteins and different molecules in a guided fashion. Additionally, generative AI allows scientists to accurately define the structure of known protein sequences to identify molecular/biological targets.

There are likely many more ways that Climate Tech startups can use generative AI to address global warming. We hope this blog post sparks ideas and inspires Climate Tech founders to use generative AI in new and exciting ways.

Generative AI workloads can consume large amounts of energy and cloud resources, and as with all workloads, it is essential to consider their environmental impact. It’s our collective responsibility to make sustainable use of this technology. Amazon is committed to reaching net-zero carbon by 2040. As part of this commitment, Amazon is on a path to powering its operations with 100 percent renewable energy by 2025, including AWS data centers. This has led to Amazon being the world’s largest corporate buyer of renewable energy for the last four years. AWS provides guidance to help companies optimize their generative AI workloads for environmental sustainability. It is also critical that these companies measure the impact of their use of generative AI and its contribution to the overall sustainability goals of the organization.

14 ways to show your startup some love with AWS Activate

Chalaire Miller — Wed, 14 Feb 2024 15:39:21 +0000

Ready to bring your next brilliant idea to life? AWS Activate is here to give your startup a little love with AWS credits, exclusive member-only offers, personalized guidance, and expert advice. This program was specially created to help startups like yours build and succeed.

The sweet benefits of AWS Activate

The goal of AWS Activate is to help your startup grow, scale, and succeed, while saving as much money as possible.

AWS Activate is 100% free to join
Start your AWS Activate journey with up to $100,000 in AWS credits
Create new infrastructure with easy-to-use, pre-made templates
Benefit from special discounts and deals, such as services, tools, memberships, and products
Enjoy the Activate console’s information and support, including personalized guidance, AWS credit details, and more

AWS Activate offers

It’s the month of love, and your startup deserves something sweet! Check out this lineup of swoon-worthy offers from AWS Activate Providers.

1. Deel

Deel is the all-in-one human resources (HR) platform for global teams. It helps companies simplify every aspect of managing an international workforce, from culture and onboarding, to local payroll and compliance.

Deel works compliantly for independent contractors and full-time employees in more than 150 countries. And getting set up takes just a few minutes.

Get the offer: AWS Activate members receive a free HRIS and 20% off the following Deel products: Employer of Record (EOR), Contractor Management & Shield, Global Payroll, US Payroll and PEO, and Entity Setup.

2. Amplitude

Amplitude’s leading digital analytics platform is the first unified system that fully understands the full customer journey and enables any team member to deliver exceptional experiences in real time. Improve data quality, ask next-level questions, identify the best customer experience, and integrate insights across multiple platforms to grow your business sustainably.

Get the offer: One year of free access to our Growth plan via Startup Scholarship (worth up to $100,000 in value) and access to lifetime discounts after that—one redemption per company.

3. Bubble

Bubble is a software development platform that enables anyone to build and launch web applications without writing code. Rather than specializing in different aspects of web development (i.e., building landing page or database), Bubble’s full-stack visual programming does it all.

Get the offer: Bubble is offering $2,500 in Bubble credits for the first 6 months of being a Bubble user.

4. Plaid

Faster, safer, more seamless financial experiences are within reach. From your first sign-up to your 10,000th, onboard new customers, fight fraud, move money, and make better risk decisions with Plaid.

Get the offer: New to Plaid? Get up to $30,000 in API credits—and no monthly minimums—for your first six months.

AWS Build guides

AWS Build is a self-service global program designed to help early stage startups take advantage of the AWS tech stack in order to build, improve, and launch their minimum viable product (MVP).

5. Build a web application on AWS

Use this guide to build a simple web application using AWS Amplify. Amplify provides a set of tools and services for building full-stack web apps, including hosting, authentication, databases, and more.

Get the guide

6. Build generative AI applications

This guide provides users with the opportunity to create basic artificial intelligence (AI)-powered demo apps, with an emphasis on fundamental principles. These hands-on sessions cover a range of topics, from creating chatbots to enhancing output with data retrieval techniques, generating images, and more.

Get the guide

7. Host your API

In this guide, you will learn how to use Amazon API Gateway to expose an AWS Lambda function. Once this API is accessible to all via the internet, you will then learn how to secure it using the Amazon Cognito user pool. This configuration will allow you to turn your statically hosted website into a dynamic web application by adding client-side JavaScript that makes AJAX calls to the exposed APIs.

Get the guide

8. Build event-driven, serverless applications

Looking to build event-driven application architectures that allow subscribers or target services to automatically perform work? Through hands-on practice, this guide will teach you the basics of event-driven design, how to choose the right AWS service for the job, and how to optimize for cost and performance.

Get the guide

9. Build automation with ML

Get started with our step-by-step guide, which illustrates a use case for a task-automated contact center using Amazon Connect, integrated with machine learning (ML). Plus, learn how Amazon Transcribe’s speech-to-text capabilities can help you accurately convert legal proceedings into text.

Get the guide

Additional benefits

In addition to guides and partner deals, new AWS Activate members receive a wide range of benefits.

10. Up to $100,000 in AWS credits for qualifying startups

Apply for up to $100,000 in AWS Cloud Credits and take advantage of a broad array of AWS services. Founders of early stage startups can use credits to build their MVP, while those running more established startups can use the credits to reduce cash burn and extend their runway.

11. Curated learning resources

Integrating AI? Building a machine learning model? Creating a mobile app? Our Startup Learning Center can assist you with countless common use cases.

12. Showcase startup directory

Showcase is a directory designed to help you amplify your business, debut your solution, find your peers, and catch investors’ attention. Showcase is your destination for extending your startup’s reach.

13. Access more than 30 startup accelerators

From generative AI to fintech and accelerators for underrepresented founders, refine your MVP alongside AWS Startup Solutions Architects and industry experts in a selection of global cohort-based programs.

14. Abundant expertise

Engage with the AWS Startups team of thousands of experts and support partners available in every region of the world. Institutional investors, technical gurus, former founders, and business leaders all come together at AWS Activate to offer you the guidance you need. Tap into a wealth of institutional knowledge from across the Amazon network with experts from Amazon Music, Audible, Alexa, and hundreds of other Amazon teams.

Looking to cozy up with AWS all year?

Startups.aws is the ultimate destination for AWS Activate members searching for special deals. You’ll benefit from niche tools, content, and resources all created specifically to help startups succeed throughout every stage of their journey. Begin growing your business on AWS with the help of AWS Startups.

How Timehop developed the world class ad platform, Nimbus, with support from AWS

AWS Editorial Team — Fri, 02 Feb 2024 17:39:40 +0000

We recently sat down with the Timehop and Nimbus CEO, Matt Raoul, as well as two talented team members, David Leviev, VP of Programmatic Product Development, and Mark Laczynski, Senior Cloud Architect, to discuss the obstacles the company has faced using third-party ad serving platforms that led to the in-house creation of Nimbus. They shared their challenges and revealed how they leveraged AWS solutions to optimize the development of Nimbus within Timehop, and not only improved the quality of ad servicing to their users, but also increased overall ad revenue.

From memories to monetization

Founded in February 2011 in a marathon Foursquare Hackathon, Timehop is an app whose primary goal is to help people connect with each other around the past. Timehop takes your digital memories, social media posts, and all your photos, packages them up into a new form of scrapbook, and serves them back to you as a daily trip down memory lane.

The idea behind it and the impeccable user experience quickly made Timehop go big. However, the rise in popularity came with its own set of obstacles. As the number of app users grew, so did the requirement for increased data storage, and processing power. The team also needed to find ways to grow the monetization model and make it more efficient. While Timehop collectively agreed that users shouldn’t be spending their dollars on their own memories, it was decided that integrating in-app ads was preferable to implementing a subscription model. But, it could not come at the expense of the user experience when viewing memories. Third-party solutions saw initial growth in ad revenue, but issues impacting the user experience, like ads taking over the user screens and unexpectedly playing sound, quickly became deal-breakers.

It became clear that the best way to significantly grow revenue without interfering with the in-app experience was to create their own in-house ad auction platform, Nimbus.

Enter Nimbus

Utilizing their experience on the AWS platform from the initial creation of Timehop, the development team turned to AWS to create the Nimbus ad platform that serves programmatic ads, gives ad blocking control, and maximizes the companies’ CPM.

“By leveraging AWS, Timehop was able to pivot its business and generate a new revenue stream by productizing Nimbus very quickly” – David Leviev

In eliminating third-party vendors from the transactions and allowing Timehop to conduct their own auctions, Nimbus proved to be a risk worth taking. After slowly rolling out the Nimbus platform to the full user base, results proved it to be more efficient, profitable, and most importantly, did not negatively impact the user experience.

The ability to connect directly to demand sources meant that the incredibly fast header-bidding auctions were no longer a pipe dream. This disruptive concept has led to higher CPM and better data transparency for Timehop.

Integrating an impression tracking system was another step forward for Timehop’s Nimbus, making real-time reporting available to publisher’s monetization teams. Letting Nimbus power the programmatic ads on Timehop resulted in operational stability and a sustained increase in profitability.

The massive success of Nimbus within the ecosphere of Timehop inspired the team to launch Nimbus as a software-as-a-service (SaaS), mobile ad-auction platform. It meant being able to help other publishers generate higher revenues with the same attention on the quality of the user’s experience.

Increasing growth with AWS Glue

Providing the tools to forecast and support Timehop and Nimbus’s infrastructure, AWS played a significant role in the company’s foundation. AWS Glue and Amazon Athena are two of the services that were most instrumental in optimizing Nimbus’ overall cost and upscaling its value.

“By implementing file conversion Glue jobs, Nimbus saw improved Athena query performance from 15 to 30 minutes down to 20 seconds, while providing a significant reduction in Athena costs and consistent and predictable costs with AWS Glue” – Mark Laczynski

Assisting with a data repository, billing, and scaling newly acquired clients, AWS provided a new level of flexibility that helped Timehop and Nimbus comprehend its value and accelerate the company’s growth.

Nimbus’ decision to build an in-house platform to meet its needs was a big leap of faith which thankfully, with a little help from AWS, turned out to be 100% worthwhile.

How Clarity AI uses AI & machine learning on AWS to quantify sustainability

Megan Crowley — Tue, 05 Dec 2023 18:06:33 +0000

Building a sustainable company—from an environmental, social, and governance perspective—is a priority for many founders. But did you know a company’s sustainability is becoming increasingly important to investors as well?

Global ESG fund assets—investments made based on how companies perform on environmental, social, and governance measures—jumped 12% in the fourth Quarter of 2022 and were on track to reach about $2.5 trillion by the end of the year. With investors and executives increasingly factoring the environmental, social, and corporate governance impacts of organizations into their decision-making, using data to track and predict these impacts has become vital.

Today, Clarity AI offers a platform built on Amazon Web Services (AWS) that provides clear and actionable data on more than 70,000 companies, 360,000 funds, 198 countries, and 199 local governments for factors such as:Impact on people and the planet through the lenses of the United Nations Sustainable Development Goals

Risk as linked to sustainability, and often through the industry consensus ESG framework supported by the standard of Sustainability Accounting Standards Board (SASB)
Climate impact measured by carbon emissions and footprint, temperature, Net Zero alignment, and TCFD reporting (Task Force on Climate-Related Financial Disclosures) powered by CDP data
Regulatory compliance (including the Sustainable Finance Disclosure Regulation (SFDR), EU taxonomy, and more)
Biodiversity impact of the investments (including the Taskforce on Nature-related Financial Disclosures (TNFD) reporting requirements) in partnership with GIST Impact

The platform provides a one-stop-shop that covers all use cases of the investors: from screening of new investments, alignment of their sustainability mandates, portfolio monitoring and rebalancing, and automatic and customized reporting to inform internal or external stakeholders.

All about analytics

How does “sustainability” of a company translate into numbers? With millions of potential data points on any given company, compiling and analyzing the metrics that are most significant can be a daunting task. Clarity AI uses artificial intelligence (including machine learning) and AWS to radically streamline and improve the process of measuring the impact of investments and organizations.

Until recent years, there were ambiguous standards of sustainability reporting which made it difficult to know what data to incorporate and how best to interpret it, explains Clarity AI’s Board Director and Vice President of Product, Ángel Agudo. “It was difficult because people were making ratings about how good or bad the companies were, without providing the ability to understand exactly what that was based on,” he says.

Building their tech stack on AWS

To provide a full view of a company’s sustainability footprint, you need to store and analyze a lot of data. To build an innovative product that does both of these things, Clarity AI decided to build with AWS.

“We try to build on AWS managed services as much as possible because they allow us to focus on optimizing the time to market while minimizing the operational costs” explains Ángel.

From a technical perspective, Clarity AI’s platform is divided into three sections:

The first is the foundational platform, which includes the cloud services, base services, and experimentation and development environments. Clarity manages their microservices using Amazon Elastic Kubernetes Service (Amazon EKS), which allows their platform engineering team to offload the availability and scalability of the Kubernetes control plane. The entry point to this set of microservices happens through Elastic Load Balancing (ELB), specifically Application Load Balancer. Storage services include Amazon Simple Storage Service (Amazon S3) and Amazon Elastic File System (Amazon EFS). Clarity AI uses purpose-built databases such as Amazon Aurora, Amazon ElastiCache, and AWS partner MongoDB. Clarity AI matches each database technology to the right use case to provide the best experience to their end users.
The next section is the data platform, which captures data and transforms it so that AI algorithms can be applied. Clarity AI orchestrates these processes with Apache Airflow, using a combination of Python ETL jobs with Amazon EMR clusters and Amazon SageMaker. Amazon S3 is the core storage technology.
The last section is the customer-facing software as a service (SaaS), which includes all of the applications that are directly exposed to Clarity AI’s customers.

Artificial intelligence (and machine learning) are key

Clarity AI’s algorithms pull data from official company sources such as sustainability reports, financial reports, and earnings calls, as well as research documents and other external sources. The platform can even parse unstructured data such as satellite images, allowing it to analyze the physical footprint of a company’s operations.

In addition, its large language models (LLMs) use natural language processing (NLP) to parse millions of news articles for new information and trends.

“We are evaluating something like 1.4 million pieces of news from more than thirty thousand trusted sources every day,” Ángel explains. “These are sources that might be capturing issues that the companies are not going to report directly, but they can say a lot about how these companies are doing from a sustainability perspective.”

Parsing the news automatically allows them to not only find new data points, but potential controversies as well. Using different NLP models, Clarity AI is able to extract the information, identify problems, and then evaluate the importance of the issues.

“A company might be named in the news for whatever reason. What we need to know is: Is this anything that might be putting the company at risk due to sustainability concerns? What is the specific issue?” says Ángel. This combination of data and contextualization gives Clarity AI an unparalleled view of exactly how a company is performing.

Automating this process allows for near-instant access to the very latest data via Clarity AI’s API or web application. If a new earnings report or news article is released, it’s incorporated immediately. AWS solutions such as SageMaker are instrumental in making this happen.

“As we train proprietary NLPs, being able to run jobs with AWS GPUs while paying only for the compute time is crucial,” says Ángel. “These NLP models enable us to efficiently collect the freshest quality data for the widest universe. We use SageMaker batch transform jobs to scalably process the inference of large numbers of documents and articles.”

Using SageMaker, Clarity AI is able to process large data analytics and AI/ML workloads quickly and efficiently. This provides the most comprehensive and granular analysis possible, be it on an asset manager’s entire portfolio or a founder’s individual company.

This is only the beginning—as the Clarity AI team continues to innovate with how AI/ML can advance sustainability initiatives, they’ve also leveraged Amazon SageMaker Studio. This fully integrated development environment (IDE) hosts Jupyter notebooks that facilitate the team’s experimentation and collaboration.

Better together

Along with meeting Clarity AI’s technical needs, collaboration with AWS has helped Clarity AI to accelerate their business: Per Ángel, “The AWS Activate program for startups has allowed us to accelerate experimentation and improve our time to market for new features.”

As a member of AWS Activate, Clarity AI has participated in training and advisory sessions with AWS experts. “This has allowed us to introduce new services and capabilities with very little friction,” says Ángel. “At the same time, AWS Activate has provided us with credits to carry out all of this experimentation with no additional cost.” He explains, “These sessions have been crucial in introducing data science services, such as SageMaker, as well as in improving our operations and security.”

AWS has also helped Clarity AI to use the AWS Well-Architected Framework to evaluate their infrastructure. “Thanks to AWS, we’ve reinforced some decisions and created new initiatives within the engineering team to improve our platform,” says Ángel. “Now, we’ve obtained ISO 27001 and SOC2 certifications and those sessions have proved extremely useful.”

Together, Clarity AI and AWS are amplifying the importance of sustainability: Clarity AI recently presented “Sustainability Assessment Powered by Machine Learning,” with support from AWS and NVIDIA, to attendees at Money20/20 in Las Vegas, Nevada. The two companies also co-hosted an ESG Opportunities panel at BattleFin’s Discovery Day New York event.

The future of better business choices

As sustainability has become increasingly valuable, investors and executives are looking for help navigating this fast-changing world. Clarity AI meets this need by using the AWS Marketplace as part of their go to market, offering their ESG risk scores API directly to companies, funds, and portfolios.

“There’s a strong relationship between AWS and financial institutions, so the more that we can use AWS to market to these institutions, the better,” says Ángel.

And it’s not just investors who are focusing on ESG issues; regulatory agencies around the world are preparing new rules on corporate environmental and social policies. As they do, a highly engaged public is paying closer attention than ever before.

“What we see on the horizon are strong and specific regulations that will force companies and financial institutions to be held accountable,” predicts Ángel. Accountability means greater access to more accurate data, which Clarity AI is ready for.

Celebrate with 12 days of AWS Activate

Chalaire Miller — Mon, 04 Dec 2023 20:13:26 +0000

The holidays are almost here. We all know what that means … presents! This winter, we’re giving you the gift that keeps on giving—AWS Activate.

AWS Activate helps startups like yours grow and flourish, including AWS Activate credits and AWS Support credits. exclusive member-only offers, tailored guidance from AWS experts, and more. You may not be able to wrap it or put it under a tree—but it’s 100% free!

The benefits of AWS Activate for your startup

AWS Activate helps your startup grow, scale affordably, and succeed while spending as little money as possible:

Join for free—new members receive up to $100,000 in AWS credits.
Receive expert technical and architecture support with up to $10,000 in AWS Support credits.
Build new infrastructure in the blink of an eye with handy pre-made templates.
Get special discounts and deals on things like services, tools, memberships, and free products.
Access the AWS Activate Console, which offers information and support, including tailored guidance, details regarding your AWS credits, and much more.

Just like ugly sweaters, one size doesn’t fit all—which is why we offer multiple membership tiers for AWS Activate. The Founders tier is for self-funded or funded (pre-series B) early stage startups, and the Portfolio tier is for startups already associated with an Activate Provider.

AWS Activate offers

Over the next 12 days, we’ll be decking the halls with deals and offers. Check back every day to unwrap a new gift handpicked just for you from the AWS Activate Provider network—hundreds of accelerators, angel investors, seed/venture capital firms, and other startup-enabling organizations. Let the savings snowball!

Day 1 – Asana

Asana allows users to manage formerly time-consuming daily tasks and key assignments all from a single, user-friendly platform. With everything safely stored in one place, you don’t have to worry about overlooking an important task or forgetting a critical job on your to-do list.

Get the offer: AWS Activate members receive one year of Asana at no cost. Campaign discount code: a4s-aws-holiday-12-free.

Day 2 – Confluent

Confluent is a cloud streaming platform based on Apache Kafka that is used for building real-time data pipelines and streaming applications. Confluent for Startups allows you to deploy the power of Apache Kafka while maintaining the speed and agility that early-stage startups need. The program received over 350 applications from startups from 47 countries and distributed $1 million in Confluent Cloud credits for free—all in its first year!

Get the offer: AWS Activate members receive up to $20,000 in Confluent Cloud credits.

Day 3 – QuickNode

QuickNode is building infrastructure and tooling to support the future of web3. Since 2017, they’ve worked with hundreds of thousands of top developers and companies, helping scale apps and providing high-performance access to 25+ blockchains.

Get the offer: Receive half a year—up to $5,000 value—of QuickNode for free.

Day 4 – Zendesk

Zendesk is a cloud-based software solution that consolidates various streams of communication—including phone, chat, and email—improving customer service and helping startups create stronger ties with their customers. Ali Rayl, Slack’s VP of Global Customer Experience, says, “I knew Zendesk Support would work if we reached the scale we were aiming for, and I also knew that the Zendesk API would allow us to build whatever we needed to meet our specific needs.”

Get the offer: AWS Activate members try six months free.

Day 5 – Vercel

Vercel’s frontend cloud provides the developer experience and infrastructure to build, scale, and secure a faster, more personalized web.

Get the offer: As an AWS Activate member, receive $1,200 worth of Vercel startup credits.

Day 6 – Slack

Slack is an AI-powered platform that increases team productivity by streamlining communication—helping team members stay on the same page, share easily, and work more effectively.

Get the offer: For one year, save 30% off eligible plan upgrades.

Day 7 – Miro

Miro is a virtual visual workspace that provides enterprise-grade infrastructure and security to help team members collaborate, communicate, and manage projects with ease—no matter where they are located. With Miro for AWS, teams can work together to build architectures with advanced diagramming tools and features like AWS Shape Packs, Reference Architecture Templates, and a Well-Architected Tool integration.

Get the offer: $1,000 credit for startups.

Day 8 – MongoDB

MongoDB is a document-oriented database program that helps developers build scalable applications with different types of data. Build the next generation of intelligent applications with MongoDB Atlas on AWS.

Get the offer: Those who sign up for a MongoDB Founders Title offer will receive up to $500 in MongoDB Cloud credits. Those who sign up for a MongoDB Portfolio Title offer will receive up to $5,000 in MongoDB cloud credits.

Day 9 – HubSpot

HubSpot offers tools to help startups find—and keep—new customers. They also offer a slew of resources intrinsic to growing your startup, including growth masterclasses, help with onboarding, software deals, and other perks.

Get the offer: As an AWS Activate member, benefit from up to 90% off on HubSpot.

Day 10 – Notion

Notion is a Connected Workspace where modern teams create and share docs, take notes, manage projects, and organize knowledge—all in one place.

Get the offer: Eligible AWS Activate members receive six months of Notion Plus for free, including Unlimited AI. Partner key: STARTUP4110P76480.

Day 11 – Intercom

Personalization is the name of the game for Intercom, which offers startups a unique and affordable customer service solution that combines an AI chatbot and a help desk with proactive support.

Get the offer: AWS Activate members benefit from Intercom’s AI-Powered Support Platform. This breaks down to 100% off in the 1st year, 50% off in the 2nd year, and a 20% off ongoing discount—up to $25,000 value.

Day 12 – Datadog

Datadog helps startups achieve better results by monitoring the different components of their cloud-scale applications, including tools, servers, containers, databases, and third-party services.

Get the offer: Benefit from up to $100,000 value with the Datadog Monitoring, Observability, and Security Platform. To qualify, you cannot be a current Datadog customer, you must join via AWS Activate, and you must have under $1.5 million in total funding.

AWS Activate offers year-round

The AWS Activate Console is the one-stop shop for special deals for AWS Activate members, with specially tailored tools, content, and resources all created specifically to help startups excel during every stage of their journey. Use the AWS Activate Console to get started on AWS and begin scaling your business.

If you’re looking to share an AWS Activate exclusive offer from your startup, check out how to become an AWS Activate Provider.

How Snorkel AI achieved over 40% cost savings by scaling machine learning workloads using Amazon EKS

Sam Huang — Tue, 21 Nov 2023 15:59:53 +0000

Machine learning (ML) startups are often heavy compute users, because they train large models using high-end GPUs and deploy them at scale for inference. AWS Startups partners with startups from inception to IPO, and has helped thousands of founders and artificial intelligence (AI) innovators build their businesses on Amazon Elastic Kubernetes Service (Amazon EKS). Amazon EKS is a popular choice to build and host ML models because it provides the flexibility of Kubernetes with the security and resiliency of being an AWS managed service that is optimized for building highly available containerized workloads.

Snorkel AI is one such company that benefits from Amazon EKS. Snorkel AI equips Fortune 500 companies, federal agencies, and AI innovators to build, adapt, and distill foundation models (FMs) and large language models (LLMs) to perform with high accuracy on domain-specific datasets. Using Snorkel’s data-centric approach toward AI development, organizations have built production-ready AI services for use cases including insurance claims processing, financial spreading, clinical trial analytics, and accelerating proactive well management for offshore drilling.

Over the past several months, the Snorkel team has been hard at work addressing the unique challenges of designing efficient infrastructure to support ML development workloads without increasing infrastructure bills, lowering developer velocity, or impairing user experience. Their ultimate objective was to reduce the cluster compute expenditure for Snorkel Flow, their end-to-end ML platform, by more than 40%.

An overview of Snorkel Flow

Snorkel’s Snorkel Flow AI data development platform enables data teams to rapidly build AI applications by employing an iterative loop of programmatic labeling, quick model training, and error analysis. Each project starts when users create a small number of labeling functions.

Labeling functions employ simple heuristics, external databases, legacy models, or even calls to large language models to apply labels to swaths of unlabeled data based on encoded expert intuition. The platform’s weak supervision algorithm combines these rule-based functions to determine the most likely label for each record. Users then train a simple model based on these probabilistic data points and assess the impact of each labeling function. In the analysis phase, users investigate slices of the data where the model underperforms. Then they build or modify labeling functions, train another quick model, and continue the loop. When users are satisfied with the quality of their labels, they build a final model on an architecture from the model zoo–ranging from logistic regression to FMs–and export it for deployment.

Due to the nature of this workflow, Snorkel Flow’s infrastructure experiences varied periods of high compute usage. Operating costs naturally increased as the customer base and ML product capabilities of Snorkel Flow scaled. To achieve efficient growth, Snorkel aimed to understand how to enhance margins while operating state-of-the-art ML software. Snorkel implemented the following practices to achieve more than 40% reduction in cluster compute costs.

Solutions for cloud cost optimization

Software-as-a-service (SaaS) startups often have opportunities to optimize their cloud spending. It is essential to understand the unique factors that drive these costs.

For Snorkel, there were two significant factors:

ML development workloads often require specialized and expensive hardware, such as GPUs. Typically, these workloads are “bursty” in nature.
Fortune 500 companies and large federal agencies use Snorkel, including major financial institutions with sophisticated IT departments that have specific deployment and data privacy requirements, by using a containerized platform.

Snorkel’s team is keen on creating systems that allow efficient scaling without a linear increase in infrastructure costs. Consequently, Snorkel developed a comprehensive autoscaling solution tailored for their ML workloads on Amazon EKS to address cloud expense concerns. This solution not only expedited workloads that require burst compute but also achieved their cost reduction goals.

In addition to the autoscaling solution, key strategies that contributed to the over 40% reduction in cloud expenses include:

Collaborating with engineering leaders and the AWS team to adopt Savings Plans through cloud configuration optimizations.
Right-sizing resources by monitoring node utilization with Prometheus and consulting backend engineers to gauge platform component needs.
Switching to cost-effective virtual machine (VM) types on Amazon EKS and utilizing multi-GPU Amazon Elastic Compute Cloud (Amazon EC2) instances for enhanced price performance.
Instituting internal process modifications where engineers collaborated with customer-facing teams to minimize idle compute.

In this post, Snorkel shares the process for addressing these scaling challenges to help facilitate the design of better infrastructure for ML systems. If you’re new to Kubernetes, read Snorkel’s Introduction to Kubernetes post to learn more about the basics and their Machine learning on Kubernetes: wisdom learned at Snorkel post to learn more about their journey with Kubernetes thus far.

What Snorkel Flow looks like on AWS

In practice, Snorkel Flow’s interaction with AWS follows the sequence outlined as follows. As the Snorkel Flow platform relies heavily on containers, the migration to AWS was almost seamless.

Users reach their Snorkel Flow instance via their web browser, which maps to a rule in Amazon Route 53.
Route 53 forwards the request to an Application Load Balancer.
The Application Load Balancer then forwards the request to Snorkel Flow pods running on a shared EKS cluster. Snorkel switched the EC2 instance type from m5 to m6a to optimize costs, resulting in a 10% compute savings with negligible performance impact based on cost per hour for the same CPU and RAM.
Additionally, they upgraded from a single g4dn.8xlarge GPU instance to a g4dn.12xlarge multi-GPU instance, which allowed for serving 4x as many GPU pods.
Each Snorkel Flow instance uses an Amazon Elastic File System (Amazon EFS) volume to store files on disk.
A self-hosted Redis queue on a pod on the EC2 instance holds the incoming jobs, waiting for worker pods to pick them up.
EKS metrics are pushed to Amazon CloudWatch, and custom scripts monitor the logs for cluster performance anomalies.

This architecture (Figure 1) has yielded a stable and snappy experience for Snorkel Flow users.

Figure 1. Snorkel Flow on AWS architecture

Rethinking how to scale

Prior to the architecture described in Figure 1, early iterations of Snorkel’s infrastructure used fixed resources. Snorkel’s users shared that these bursty workloads could take too long to complete and therefore negatively impacted their experience.

Manual scaling of compute resources proved unscalable and error-prone, leading to cloud costs that stayed elevated even during periods of low usage. It was the worst of both worlds: low cloud cost efficiency and slower-than-needed performance.

To address these challenges, Snorkel implemented autoscaling at multiple levels in their infrastructure, as discussed in the following sections.

Designing scalable infrastructure with cost-efficiency in mind

The Kubernetes distribution of Snorkel Flow involves a set of deployments running in an EKS cluster that contains pods that run various components of the platform.

As shown in Figure 2, to address the unique challenges of working with bursty compute workloads, Snorkel’s team introduced a new concept for Kubernetes pods: semantically categorizing them as either “fixed” or “flexible.”

Pods that are fixed cannot be safely moved from node to node, either because they will lose important in-memory state (such as in-progress compute jobs without checkpointing) or to minimize avoidable downtime for foundational platform components (for example, the orchestrator for the Ray cluster).
Pods that are flexible can be safely moved to a new node. This distinction is meaningful in the context of autoscaling, because downscaling nodes involves moving pods away from underutilized nodes when they are terminated.

Figure 2. Examples of pods and their associated workloads under the fixed/flexible framework

This fixed/flexible framework gives Snorkel a domain-specific means to enable automated cluster downscaling, which allows them to turn on the cluster autoscaler on Amazon EKS without their finance department messaging them every hour.

Snorkel’s initial approach was to deploy podDisruptionBudgets on the EKS cluster to prevent the cluster autoscaler from moving flexible pods during the day and from moving fixed pods at all. While effective, this approach left the Snorkel team unsatisfied because it downscaled far fewer nodes than what was theoretically optimal.

To address this, Snorkel layered on a pod scheduling optimization that isolated fixed pods to a small fixed group of nodes. It scheduled flexible and worker pods (which are considered to be fixed pods but are ephemeral due to worker node autoscaling) in the remaining flexible group of nodes.

These changes allowed Snorkel to efficiently downscale the flexible nodes at night, when it became safe to move around flexible pods and scale down the vast majority of worker pods.

Enabling efficient downscaling of the vast majority of the cluster’s nodes (i.e., the flexible nodes) allowed Snorkel to meet their target of reducing the cloud costs for hosting Snorkel Flow by over 40%.

More detail on Snorkel’s autoscaling solution

Snorkel divides the implementation of the solution described in the previous section into three sequential efforts:

First, Snorkel implemented “worker autoscaling,” a custom Redis-based autoscaling service that allows their worker pods to scale up and down based on jobs in the workers’ queues.
Second, they implemented “cluster autoscaling” by reconfiguring their Kubernetes deployments to allow the Kubernetes cluster autoscaler to scale nodes down, in addition to scaling nodes up.
Third, Snorkel implemented “node downscaling optimizations” by grouping fixed pods into a small group of fixed nodes to prevent fixed pods from interfering with the downscaling of the remaining nodes.

Worker auto scaling

The Snorkel Flow platform abstracts compute into a paradigm where jobs wait in Redis queues and workers run as processes in worker pods.

Snorkel implemented a worker autoscaling solution (Figure 3) for worker pods by running a recurring function in Snorkel Flow’s backend API. Every few seconds, this function checks the Kubernetes cluster and Redis for both upscaling and downscaling eligibility.

If there are jobs waiting in one or more relevant Redis based queues, the function will ask the Kubernetes API to provision additional worker pods to process these jobs. If the Redis queue is empty and there are no running jobs in the job registry, it will ask the Kubernetes API to destroy the worker pods to free up reserved CPU and RAM resources.

Figure 3. Architecture of the worker autoscaling implementation

As shown in Figure 4, with this worker autoscaling implementation rolled out, Snorkel Flow’s worker pods became ephemeral, appearing in the cluster only when jobs needed to be processed.

Figure 4. The worker autoscaler scales the number of worker pods up and down during the day based on compute demands

Cluster autoscaling

The PodDisruptionBudget resource protects certain pods against disruption (for example, voluntary restarts) by allowing for the specification of the maximum number of pod replicas that can be unavailable at any given time. As shown in Figure 5, setting this value explicitly to 0 for a deployment ensures that the cluster autoscaler will not downscale nodes running the deployment’s pods.

Figure 5. Node 1 is underutilized and has a pod with a defined PodDisruptionBudget where maxUnavailable is 0 (red). Node 2 is underutilized and all its pods have a PodDisruptionBudget where maxUnavailable is 1 (blue). The cluster autoscaler will terminate Node 2 for underutilization, but cannot terminate Node 1 because of the red pods

Implementing this resource on hosted Snorkel Flow instances safely allowed the cluster autoscaler to downscale underutilized nodes. However, the cost savings Snorkel realized were marginal—they were still unable to downscale the majority of their nodes because all Snorkel Flow pods were protected by an associated podDisruptionBudget.

Upon closer examination, Snorkel’s team realized that this protection does not need to exist all of the time. Workloads are bursty, and most user interaction with Snorkel Flow occurs during a customer’s business day, meaning it is safe to relax this protection outside of business hours. Similar to worker autoscaling, Snorkel implemented a recurring function that toggled podDisruptionBudgets “off” overnight for an instance’s flexible pods by setting the maximum number of unavailable pod replicas to 1, up from 0 (Figure 6). The previous worker autoscaling solution combined with the ClusterAutoscaler and PodDisruptionBudget solution was able to downscale many more underutilized worker nodes than before. Customers deploying Snorkel Flow in their cloud can configure this as needed.

Figure 6. The cluster scales up nodes during the day to support normal operations during business hours, scales them down at the end of the day, and scales them back up periodically during the night for CI workloads

Node downscaling optimizations

Even with these improvements, Snorkel saw that a majority of underutilized nodes were not being downscaled at all.

Upon further investigation, Snorkel realized that the issue stemmed from fixed and flexible pods occupying the same node. This was problematic because a fixed pod, pseudo-randomly assigned to a node containing flexible pods, would “pin” that node and prevent it from being downscaled, even when it is underutilized. This lack of control over the scheduling of fixed pods led to periods where the vast majority of the cluster’s nodes could not be downscaled, even though they represented far more compute power than needed at the time.

Snorkel leveraged the podAffinities Kubernetes resource to address this, which enabled them to constrain which nodes a pod is eligible to run on based on the labels of other pods already running on any given node. They added labels to the pods to differentiate between fixed versus flexible pods, and added a podAntiAffinity stanza to their deployments configuration to ensure that fixed pods are not scheduled on nodes running flexible pods, and vice versa.

As shown in Figure 7, this implementation of podAffinities allowed Snorkel AI to split nodes into two functional groups: the fixed group of nodes containing fixed pods, which can never be safely moved between nodes (for example, Redis due to cache), and the flexible group of nodes containing “flexible” pods that are either ephemeral (such as worker pods) or safe to move outside of business hours (such as overnight).

Although it is possible with manual intervention during platform maintenance, Snorkel cannot automatically downscale the fixed nodes. This solution, however, allows them to automatically downscale the flexible nodes because they have now isolated the unmovable pods into the fixed nodes.

Figure 7. The implementation of podAffinities enabled Snorkel to move from pseudo-random pod-to-node assignment of red (fixed) pods and blue (flexible) pods (top) to a more intentional approach (bottom). Previously, the cluster autoscaler could not downscale any of the three underutilized nodes due to the presence of red pods on each node. By grouping the red pods into their own group of nodes, the cluster autoscaler effectively downscales all remaining nodes

This semantic differentiation between groups of nodes helped Snorkel realize significant additional savings, because the flexible group of nodes represents greater than 90% of the cluster’s total nodes.

Conclusion

The Snorkel and AWS Startups teams hope that sharing this thought process and these solutions helps other startups to build better infrastructure for ML workloads, which are rapidly becoming more important as ML, large language models, and other FMs make their way into production for organizations all over the world.

Attending AWS re:Invent 2023? This deployment will be highlighted as part of Snorkel’s session Navigating the future of AI: Deploying generative models on Amazon EKS (session CON312). Be sure to check it out!

Acknowledgments

Many thanks to David Hao, Edmond Liu, and Alec Xiang for helping to make this technical vision a reality for Snorkel. Special thanks to the aforementioned as well as Matt Casey, Henry Ehrenberg, Anthony Bishopric, and the entire Snorkel infrastructure engineering team for their thoughtful feedback on this article.

Crate builds with generative AI on AWS to reimagine online spaces

Megan Crowley — Mon, 13 Nov 2023 16:46:31 +0000

Previous generations filled their attics with boxes of pictures, clothing, and souvenirs as memories of their lives. Today, more and more memorabilia lives online as texts, images, videos, snippets of poetry, chat comments, and more. This media from personal corners of the internet requires an entirely new kind of storage: Spaces created not with physical tools, but through lines of code by companies like Crate.

Anna Bofa, founder and CEO

Crate, founded in 2021 by chief executive officer (CEO) Anna Bofa and built on Amazon Web Services (AWS), offers an online storage system inclusive of all forms of digital artifacts. With the Crate app, TikToks can live alongside photos, music tracks, news articles, and more. These “crates” are shareable collections of content you collect from different platforms, all organized and powered by personalized artificial intelligence (AI). The company was born of Anna’s belief that, at its core, the internet is a collection of content and people.

“In my early thinking about Crate, I wanted to provide easy opportunities for people to connect over content in joyful ways,” says Anna. “I envisioned a shared happy place away from some of the more negative aspects of social media.”

“We like to say that we see crates everywhere. For example, you can make a crate of articles that really resonate with you or a crate of all the places you want to visit. You can add recommendations from friends, and store it all in one place,” Anna explains.

For any given crate, the app generates a cover, title, and tag, while also using AI-generated language models to build a personalized understanding of a user’s preferences and interests. Anna’s ambition for Crate is to create an internet in which the user does not go out looking for content. Rather, the best of the internet’s digital assets come to them.

Proving what’s possible with generative AI

The Crate app helps users collect, connect, and share the best of the internet.

Through her experience working at Google, Facebook, and multiple startups, Anna is well aware of the power of AI to transform the internet. “What we do with AI in the early days really matters because it will be the foundation for everything else,” she explains. Using generative AI to raise the bar on Crate’s product has been a key component of the startup’s success.

The Crate team is proving it’s possible to generate content across multiple mediums–such as the text and images for crates–via one easy-to-use interface, while also personalizing media suggestions based on your saved content. For example, let’s say you’re interested in urban gardening. You save relevant digital content of any sort to your gardening crate. Based on that collection, Crate will then automatically go out and find additional gardening content for you. With Crate, the idea is that you don’t have to be a prompt engineer to gather content that matters.

Anna sees Crate as offering a solution to what she believes is the internet’s biggest problem: information overload.

“It used to be that you went online to read the morning news, check Facebook or Tumblr. Now, with everything online, the magic of the internet is also its biggest pain point because the amount of information is overwhelming and fragmented.”

Crate is working to change that by making the abundance of online content more manageable and enjoyable, simplifying the user’s experience. Or, in Anna’s words, the app is “bringing the magic back.”

Accelerating success with AWS

From the beginning, Anna and the Crate team have relied on AWS for support–whether that’s technical solutions, business expertise, or networking opportunities. “As a startup, we literally could not have built our product without AWS,” explains Anna.

Building with AWS solutions

From a technical perspective, Crate’s entire infrastructure is built on AWS. Crate’s ability to save content from anywhere across the web and organize it using AI-powered automation occurs via two distinct stages: content ingestion (pre-processing and indexing, which happens in the background after a user saves a piece of content) and downstream tasks (organizational features that are facilitated by the processes that take place during content ingestion). While Crate leverages AWS solutions for both stages, content ingestion is the most computationally intensive part of their product and an area where they’ve greatly benefitted from building with AWS services.

Content ingestion is an event-driven process facilitated by Amazon Kinesis and step functions, which entails a number of steps including: content retrieval; image captioning, performed using a model hosted in Amazon SageMaker; video transcription using Amazon Transcribe; content summarization; the generation of semantic embedding vectors, stored OpenSearch to support semantic search and clustering, and more.

Each step happens asynchronously, triggered by events which are sent and via Amazon Kinesis and made even more robust with the use of Amazon Simple Queue Service (Amazon SQS), in the case of a failure in one of the steps. Additionally, most of the steps run in Amazon Lambda functions which are cost-effective and enable efficient deployment and iteration.

Kinesis is particularly critical to Crate’s tech stack because it enables the asynchronous processing of data in a robust and decoupled manner, allowing them to add or remove steps easily. Additionally, the Crate team shares that SageMaker is also useful for deploying smaller models in a hassle-free, robust way that saves their small team much time.

“What’s beautiful about the AWS platform is that everything is already built and we don’t have to spend time on the heavy lifting,” says Anna. “AWS understands how systems come together—no matter what our need, AWS has something to meet it.”

Leveraging AWS programs

As a member of AWS Activate, a program for startups offering technical support, architecture guidance, and cloud credits, Crate has had access to the technologies and technical expertise it has needed. With AWS Activate’s technical support and credits, Crate has been able to grow to where the company can now effectively begin to monetize or raise funds. “As a founder, I’m always thinking about costs,” says Anna. “We’ve been fortunate to benefit from credits through multiple AWS programs. It’s been transformative for levelling the playing field and allowing us to build a competitive product.”

Along with AWS Activate, Crate also participated in the AWS Generative AI Accelerator, a 10-week program designed to take the most promising AI startups globally to the next level. It provides access to AI models and tools, customized business strategies, machine learning stack optimization, and more. On the business side, the accelerator connects entrepreneurs to investors, customers, experts, and networking opportunities.

“One of the most important benefits of the AWS Generative AI Accelerator was being able to talk directly with Amazon’s internal talent,” says Anna. “We had a biweekly call with AWS engineers. They helped us build things we hadn’t been able to build before.”

Anna’s goal in the accelerator program was to launch the Crate beta app and come out of stealth mode. She met that goal, launching Crate in app stores the day after the program finished. “The accelerator also introduced us to investors, which has really helped,” explains Anna.

Anna pitches to investors, press, and customers during Demo Day for the AWS Generative AI Accelerator

Envisioning the future of Crate

As the Crate team continues to seek more efficient ways to scale their product and delight their users, Anna plans to move more of the app’s stack to AWS. The Crate team has tested Amazon Bedrock, for instance–a fully managed service that offers a choice of high-performing foundation models (FMs) from leading AI companies through a single API, simplifying development while maintaining privacy and security. Crate is now in beta and available for early users to try.

Members of the team share a ‘C’ for Crate

Anna believes that the potential for AI to revolutionize how we live and work is unparalleled, and startups will lead the way. To other founders, she advises, “Dive in. Explore the potential. Participate in AI communities and discussions. Envision a positive future for AI and build toward it.”

Top 10 unforgettable moments from AWS GenAI Day

Chalaire Miller — Fri, 13 Oct 2023 14:45:37 +0000

After a whirlwind of innovation and inspiration, AWS GenAI Day has come to an end and the memories are unforgettable.

As the virtual curtains draw on this extraordinary event, we can’t help but reflect on the incredible insights, groundbreaking discussions, and cutting-edge technologies that shaped the day. From diving deep into AI-driven data augmentation to exploring the limitless potential of generative models, this journey has been nothing short of transformative.

The story doesn’t end there. Join us as we take a captivating stroll down memory lane, revisiting the highlights, key takeaways, and memorable moments that defined AWS GenAI Day.

From visionary speakers to thought-provoking sessions, we captured the essence of this event and how it’s shaping the future of AI.

So, fasten your seat belts and get ready for an exhilarating wrap-up that celebrates the power of generative AI and sets the stage for what’s to come. The adventure continues!

1. Exploring the wonders of deep fakes

Have you ever imagined a world where humans float effortlessly or dogs contemplate life’s mysteries with a touch of human wisdom? It may sound like the stuff of science fiction, but thanks to the magic of generative AI and deep fake technology, these whimsical scenarios are no longer just fantasies.

In the lead-up to AWS GenAI Day, our social platforms treated our audience to a series of captivating teasers that showcased the incredible potential of deep fakes. From the enigmatic “Brainy Floater” to the charming “Human Dog,” these characters left followers both curious and astounded.

The creative teasers not only sparked the curiosity of those exposed to them, but also highlighted the incredible potential of generative AI technology. They serve as a testament to the limitless boundaries of innovation and the power of imagination.

AWS GenAI Day features creative deep fakes as teasers to the event

2. Insights from the streets

In the run-up to AWS GenAI Day, we took to the streets to capture the pulse of curiosity and creativity in the world of generative AI. Led by AWS Senior Developer Advocate, Brooke Jamieson, these “man on the street” videos gave us a glimpse into the captivating realm of predictions, powered by the magic of generative AI.

What if you could predict the unpredictable? With generative AI at our fingertips, that tantalizing prospect is closer than ever. Brooke engaged with tech enthusiasts to unearth their thoughts on the predictions they wish to harness through this groundbreaking technology.

These insightful videos remind us of the boundless possibilities that await. The streets were buzzing with enthusiasm, and it’s clear that the world is hungry for the transformative power of generative AI.

AWS Senior Developer Advocate Brooke Jamieson seeks tech enthusiasts’ thoughts about generative AI

3. Charting new horizons

AWS GenAI Day kicked off with a resounding welcome, setting the stage for a day of exploration, innovation, and transformation. The Executive Welcome, led by Howard Wright, Vice President of Startups, and the visionary Swami Sivasubramanian, AWS VP of Databases, Analytics, and ML, was a powerful testament to the profound impact of generative AI on startups and the entrepreneurial ecosystem.

As these two formidable leaders took the stage, Howard and Swami ignited a spark of inspiration that would guide us throughout AWS GenAI Day. Their opening address was a promise—an assurance that the day ahead would be filled with insights, discoveries, and opportunities. The Executive Welcome set the tone for the entire event, inviting all of us to embark on a transformative journey. It was a journey that would delve deep into the world of AI, reshaping the landscape for startups, entrepreneurs, and visionaries alike.

Howard Wright and Swami Sivasubramanian welcome attendees to AWS GenAI Day

4. A vision for the future

The echoes of AWS GenAI Day continue to resonate as we delve into one of its most impactful sessions: the Keynote Panel discussion, “Mapping the Trajectory of GenAI: From Learning to Impact.” This pivotal event left attendees inspired and enlightened as it explored the profound influence of generative AI across diverse industries. The keynote panel served as a powerful catalyst for innovation and collaboration, envisioning a future where generative AI continues to drive positive change and shape our collective progress. The impact of this session extends beyond the confines of AWS GenAI Day. It has ignited a spark—a commitment to harness the power of generative AI for the betterment of society and a shared determination to navigate this transformative journey responsibly.

From the insights gleaned in this keynote panel, we are reminded that the generative AI revolution is not only about technology; it’s about using it to create a brighter, more inclusive, and ethically sound future. The journey has only just begun, and the possibilities are boundless.

Speakers at the Keynote Panel discuss how they envision the trajectory of generative AI

5. Redefining education

At AWS GenAI Day, a session that truly captured the spirit of innovation and transformation unfolded under the banner of “Teaching a Generation to Read: Generative AI’s Impact on Education.” This remarkable session revealed how generative AI, trained on user-generated data, is now catalyzing a revolution in the world of education.

Catalin Voss, co-founder and chief technology officer (CTO) of Ello, went beyond the transformative technology and also shared valuable insights on best practices and considerations for startups across industries. This session served as a powerful reminder of the limitless possibilities that generative AI brings to education and beyond. It’s a force that can reshape not only how we learn but also how we interact with the world around us.

Catalin Voss and Davis Stone discuss how generative AI is redefining education

6. Generative AI transforms productivity and collaboration

In a world where the mundane is automated and collaboration reaches new heights, the future of work is being rewritten. AWS GenAI Day unveiled a captivating panel discussion featuring founders from pioneering companies like Crate, Knowtex, Fireworks, and Yellow. These visionaries are harnessing the power of generative AI to make work not just productive but also enjoyable and rewarding.

The panel discussion challenged us to imagine a work landscape where professionals can focus on outcomes and passion, leaving the tedium to generative AI. It’s a vision that’s not just imagined, but actively pursued by the founders on the panel.

From automating repetitive tasks to enhancing collaboration, the founders shared how they are implementing generative AI to achieve exceptional results. It’s a testament to the transformative potential of this technology, which is reshaping the way we work and collaborate.

The discussion went beyond showcasing the founders’ success stories; it also served as an inspiration for businesses of all sizes. Generative AI is not just a tool; it’s a catalyst for positive change. It empowers businesses to reimagine their operations, putting productivity and collaboration at the forefront.

This enlightening panel discussion reminded us that generative AI is more than a technology; it’s a gateway to a more efficient, enjoyable, and rewarding future of work. The founders from Crate, Knowtex, Fireworks.ai, and Yellow.ai are pioneers in this journey, and their stories serve as beacons of inspiration for us all.

Panelists discuss how generative AI will result in the next productivity frontier

7. The power of open source in generative AI

In the fast-evolving landscape of generative AI, open-source frameworks have emerged as the driving force behind a wave of groundbreaking applications. AWS GenAI Day brought to the forefront the immense potential these frameworks hold, particularly in the realm of language model-powered applications. Harrison Chase, the co-founder and chief executive officer (CEO) of LangChain, shared invaluable insights into harnessing the power of this revolutionary framework on AWS.

The session underscored the pivotal role that open-source frameworks play in fueling innovation in the generative AI domain. Businesses and developers alike are realizing the vast possibilities of these frameworks in building cutting-edge applications.

Harrison Chase shares his insights on building and deploying cutting-edge generative AI applications

8. Pioneering in speech recognition

With demand for sophisticated models and speech recognition technology skyrocketing, tackling the unique technical challenges of large-scale AI model inference becomes paramount. AWS GenAI Day provided a platform for experts like Domenic Donato, Vice President (VP) of Technology at AssemblyAI, to share insights into optimizing computational resources while delivering top-notch performance.

Domenic Donato delved into the heart of the matter: the unique technical challenges associated with large-scale AI model inference. As the demand for AI-driven speech recognition soars, optimizing computational resources becomes a critical consideration.

The session provided invaluable insights into how AssemblyAI addresses this challenge. From fine-tuning models to efficient resource allocation, Domenic shed light on the strategies and best practices that ensure optimal performance at scale.

AssemblyAI and AWS discuss techniques for effective and efficient large-scale inference

9. Shaping generative AI investment strategies

This enlightening panel discussion dove deep into the world of venture capital and its dynamic relationship with generative AI. The engaging conversation brought together a diverse group of investors who are leading the way in funding AI-driven innovations.

Investing in AI-based ventures comes with unique challenges and opportunities. The panel explored the intricacies of this landscape, highlighting the transformative potential of AI across various industries.

The panel featured industry experts such as Sarah Guo, Jon Turow, and Kleida Martiro, whose collective insights provided a deeper understanding of AI and the venture capital landscape. Their expertise resonated with entrepreneurs seeking funding, investors looking to navigate the AI investment landscape, and anyone interested in staying informed about cutting-edge developments.

Investors and AWS discuss how startups can navigate the AI investment frontier

10. OctoAI to the rescue

As AWS GenAI Day reached its pinnacle, a standout solution took the stage, promising to democratize the world of generative AI and make it more accessible and cost-effective. OctoAI compute service, a developer platform designed for deploying and customizing generative AI models, held the spotlight.

The session was enriched by the presence of experts in the field. Luis Ceze, CEO and co-founder of OctoML; Vanessa Yan, Staff Product Manager at OctoML; and Deepam Mishra, Senior Advisor in AI/ML Startups and Venture Capital at AWS, shared their insights into how OctoML and AWS are collaborating to make generative AI work reliably, efficiently, and sustainably.

In this game-changing session, we were reminded that the power of generative AI is now within reach for developers and businesses of all sizes. OctoAI compute service is a testament to how innovation and collaboration can drive accessibility and cost-effectiveness in the world of AI.

A discussion on making generative AI work reliably, efficiently and sustainably with OctoML

The future of generative AI is brighter than ever, thanks to solutions like OctoAI. AWS GenAI Day continues to be a source of inspiration and empowerment, paving the way for a more inclusive and innovative AI landscape.

As we reflect on a day filled with groundbreaking insights, transformative technologies, and a community of innovators poised to reshape the future, these unforgettable moments from AWS GenAI Day encapsulate the essence of the day—a journey of innovation and exploration into the limitless potential of generative AI. The future of technology is here, and it’s powered by generative AI.

Want to see more?

Check out all of the memorable moments on the AWS Startups social channels:

AWS Startups X for the timely and behind-the-scenes content
AWS Startups LinkedIn for the latest insights
AWS Startups Facebook for the biggest news
AWS Instagram for the motivational and inspirational

VC perspectives on generative AI for Latino startups

Nehemiah Green — Mon, 09 Oct 2023 13:21:42 +0000

The strategic decisions made early in a startup’s journey often play a large role in the outcome of the business. These decisions are even more crucial for founders who are underrepresented in their industry; they may not have access to the same resources to sustain many pivots and iterations of their core business strategy. In particular, we’d like to highlight that in the unique investment cycles in artificial intelligence (AI) and machine learning (ML), investors are becoming more stringent. Ensuring that business fundamentals are solid is of the utmost importance, especially considering the historical underinvestment in Latino founders.

The good news is that we are living in a moment of extraordinary potential. Generative AI has already changed the game for startup founders, lowering the time it takes to achieve product viability, and opening the gates of technology development to founders with fewer resources and smaller networks. At Amazon Web Services (AWS), we’re committed to helping startups take full advantage of these opportunities through our Underrepresented Founder Startup Business Development Team, which works with women founders, founders of color, founders with disabilities, and LGBTQIA+ founders to accelerate their progress. Opportune moments like this one happen once in a generation, and it’s up to us to seize it.

Through the Underrepresented Founder/Investor Business Development program at AWS, we had the great privilege of partnering with SomosVC, an extraordinary group of Latino investors, for a panel discussion on building generative AI businesses by and for the Latin American community. This brings us back to the idea we started with: How can underrepresented founders incorporate generative AI into their startup?

Samuel Garcia, a partner at Amplo VC

Even in an early-stage market such as generative AI, the barriers to entry can be daunting. Not only do the incumbents have brand recognition and capital to deploy, they also have a wealth of proprietary data built up over years of customer interactions—a key ingredient to building meaningful AI products.

Just because they have the data, though, doesn’t mean they know how to use it, argues Samuel Garcia, a partner at Amplo VC. Step one for companies trying to compete with larger players is to build or obtain a proprietary dataset that’s going to be difficult for competitors to copy. Put that together with a team of dedicated AI and machine learning engineers, and “those folks can really rival these incumbents, because they can do a lot more than the incumbents can,” Samuel says.

Alan Arturo Day, an investor at New York Life Ventures

How? By focusing on the hard problems. Fortunately for ambitious founders, established companies have an excess of data that startups can help unlock value from, says Alan Arturo Day, an investor with New York Life Ventures. Initial use cases are likely to focus on internally-facing workflows which represent immediate opportunity for startups to validate their technology. “It’s not as high-risk and there’s not as many regulatory hurdles,” Alan says.

Other use cases for generative AI are further out on the horizon, and these higher-stakes, often customer-facing applications—whether it’s helping customers decide on the right insurance plan or giving them investment advice— will take more testing, implementation, and time to drive enterprise adoption.

Lisette Tellez, Managing Director of Ocean Azul Partners

The regulatory hurdles that might deter a large company can also play to a startup’s advantage—if the team is willing to pivot. Lisette Tellez is the President of the Board of SomosVC and the Managing Director of Ocean Azul Partners. She sees potential in companies that have done the hard work of gaining certifications and clearances and now just need the right product to take full advantage of the generative AI moment.

“One of our portfolio companies is in a regulated space—banking—and they’ve found that’s a really great moat for them,” Lisette says. Newer startups working to offer AI chat services to financial companies pose little competition because the older company has already established procurement processes with their existing clients. “For existing companies to add a new product line or feature is just going to make that product all the more sticky,” she explains.

This isn’t just about adding on “generative AI wrappers that don’t really bring a lot of differentiation to the table,” however, as Samuel puts it. It’s about companies harnessing AI to set themselves apart. When he’s considering whether to invest in a particular AI product, Samuel explains he’s looking not just for companies with a wealth of proprietary data, but also ones that are designing products with revenue in mind. “A whole other product offering that we have this data set for—that’s a no-brainer,” he says. “But when it comes to, ‘Should we add this feature?’ it really comes down to, ‘Can we charge more for it?’”

Sergio Monsalve, Founding Partner of Roble Ventures

The other key question is, of course, is ‘What is the competition?’ And equally importantly, how are they doing it? Sergio Monsalve, Founding Partner of Roble Ventures, has been thinking about the relative advantage of software as a service (SaaS) versus managed service models. In this transition moment from the era of big data into the era of generative AI, he says, “I’m not sure the set-it-and-forget-it SaaS model is going to work.” Instead, “it makes a lot more sense for startups to grab the data, manage it for the customer, give them the answer, and then bill a lot of money for that service.”

The best way to compete in the generative AI space may be to join forces: merge, roll up, and acquire. “AI is pretty much like one plus one equals three,” Sergio adds. “If you put two companies together that have proprietary datasets with some very good machine learning (ML) and some really good AI people, then you’re going to get the synergies to get real return on investment.”

As a founder looking to get acquired, it’s important to have a clean and efficient tech stack, Lisette adds. But even better is to lean into what makes you unique. For Latin American founders, that might mean tailoring products where you have a deep expertise and in partnership with resources such as SomosVC and FounderFamilia. If you optimize your product to a customer segment’s specific needs, you create a data flywheel—the more customers you attract, the better and more refined your product becomes.

There’s also an incredible opportunity to tap into a pool of talented AI and ML engineers. At AWS In Latin America, we’ve observed startup founders, such as CEO Israel Niezen who co-founded Factored, leverage near-shoring. This approach has not only offered exceptional caliber of talent in US time zones, but has also delivered a cost-effective solution, diversified his team, and supported his community. As always, both SomosVC and AWS are here to assist.

Mariela Salas, Executive Director at SomosVC

“I want to point out one thing,” SomosVC Executive Director Mariela Salas says at the end of the discussion. “The fact that we have a panel of five Latinos talking about AI is amazing. I don’t know that we’ve ever seen this before. I’m excited to see it happen more often.”

According to the SomosVC 2022 Annual Report, Latino investors remain 2% of the industry, relative to 19% of the U.S. population. “Our organization is focused on growing and supporting the Latino/a venture capital ecosystem,” explains Mariela. “Specifically, we aim to increase the representation of Latino/a professionals in venture capital, help Latino/a venture investors build their careers and networks, and improve access to capital for Latino/a-led venture firms.”

AWS, along with LatinoVC and other organizations, are hosting candid dialogues to bridge the gap between Latino investors and founders. These discussions equip founders with insights to hone their fundraising strategies, boosting their potential for success. Furthermore, these dialogues guide founders in aligning their product development with the priorities and needs that investors deem crucial. We’re committed to continuing this initiative. Through partnerships with VCs, universities, incubators, accelerators, co-working spaces—you name it—we’re working to make sure founders from Latin American and other communities continue to rise and thrive.

Have questions or ideas? Now’s the time to reach out: nehgre@amazon.com to reach Nehemiah and the Underrepresented Founder Startup Business Development Team. We want to hear from you.

Ready to begin your startup journey? Join AWS Activate to build and scale your startup with the right resources at the right time.

Learn how underrepresented founders are proving what’s possible with AI on AWS:

AWS Activate updates program benefits regularly, and credit offerings and/or the offerings reflected in this blog post may differ from current Activate offers. For the most up to date information about Activate benefits, please visit https://aws.amazon.com/activate/

AWS Startups declares week of November 6th Women’s Demo Week

Charlotte Newman — Wed, 04 Oct 2023 11:56:50 +0000

AWS Startups is excited to announce the inauguration of a global movement, AWS Startups Women’s Demo Week, set to take place annually from November 6th – 10th. Designed to honor and elevate the incredible contributions of women founders worldwide, this movement is a testament to AWS Startups’ unwavering commitment to diversity, inclusivity, and the transformative power of innovation. This week culminates a years-long series of events designed to engage and accelerate the startup journeys of women founders around the world, including AWS Founders Forums and roundtables in India, Indonesia, Kenya, Malaysia, Singapore, and Vietnam connecting women founders to AWS experts, investors, and technical resources.

There’s no denying the crucial role that women play in the global startup ecosystem. However, despite their enormous potential, women-led startups often encounter unique hurdles in acquiring resources, opportunities for networking, and capital. This is why AWS launched the Underrepresented Founder and Investor Startup Business Development team in 2020 to lead a global strategy to address the underinvestment in women and other diverse founders. In 2020, only 2.3% of global venture capital (VC) funding went to women-only founded startups, with Black and Latina women receiving less than 1% each. Unfortunately, this funding gap worsened in 2022—the latest data from Crunchbase shows VC funding for all-women founding teams dropped even further—to just 1.9% last year, with women of color receiving less than .35%. These stark statistics spotlight the sizable and growing funding discrepancies facing women entrepreneurs worldwide, especially women of color. They underscore why groundbreaking initiatives like Women’s Demo Week are so critical.

Our mission with Women’s Demo Week is to create an impactful, inclusive space that champions and strengthens the role of women in the startup ecosystem. By providing a platform for showcasing groundbreaking innovations and fostering connections with investors and industry stakeholders, we aim to catalyze change and progress towards a more equitable startup ecosystem. Throughout Women’s Demo Week, AWS will host demo days worldwide in New York, San Francisco, Washington DC, Los Angeles, and Austin in North America; Bogota, Buenos Aires, Santiago, Sao Paolo, and Mexico City in Latin America; and London, Berlin, Lagos, Johannesburg, Cape Town, Tel Aviv, Amsterdam, and Madrid in Europe, the Middle East, and Africa.

This event offers selected entrepreneurs from our AWS Startup community the opportunity to connect with potential investors, customers, strategic partners, and other industry influencers. The agenda for the event goes beyond just showcasing innovative solutions—it encompasses sharing visionary ideas, building stronger networks, and celebrating the women entrepreneurs who are reshaping industries to become more diverse. We invite you to join us—whether you’re an investor seeking promising startups, a potential customer keen to explore innovative solutions, or an entrepreneur looking for inspiration, Women’s Demo Week will be a week filled with learning, networking, and celebrating the indomitable spirit of women entrepreneurs. Register today and learn more!

Sign-Speak builds with AI on AWS to create accessible experiences

Megan Crowley — Mon, 25 Sep 2023 13:43:10 +0000

If you were to encounter one of the 430 million Deaf and Hard of Hearing (D/HH) individuals in the world, there are two main ways you can communicate with them: interpreters or texting back and forth. Interpreters, while providing great services, often are cost prohibitive for certain situations and texting is not the most optimal. Sign-Speak is changing that paradigm.

Sign-Speak—a startup whose language software recognizes American Sign Language (ASL) and translates it into spoken words (and vice versa) with machine learning—is doing exactly that. With the World Health Organization (WHO) predicting that nearly 1 in 10 people will have disabling hearing loss by 2050, Sign-Speak is building a solution to improve lives today and well into the future.

Founded in 2020 by chief executive officer (CEO) Yami Payano, chief product officer (CPO) Nikolas Kelly, and chief technology officer (CTO) Nicholas Wilkins, Sign-Speak is proving what’s possible when accessibility and machine learning (ML) merge. All three founders have a foundation in ASL, with a passion for building with (not for) the Deaf community that comes from personal experience: Nikolas, a Deaf individual, has experienced the challenges that arise from accessibility barriers. Yami has a family friend (like family) who—as a result of being Deaf—was unable to attend school in the Dominican Republic because hiring an interpreter was more expensive than the entire cost of tuition.

“From an early age, I grew up understanding the lack of access,” explains Yami. “As the child of Dominican parents who cannot communicate in English, I empathize with the children of Deaf parents. Like me, they often have the role of translator. No individual should face that kind of responsibility.”

Yami Payano, CEO Nikolas Kelly, CPO Nicholas Wilkins, CTO

To make the world a more accessible and inclusive place, the Sign-Speak team built a platform that offers real-time ASL recognition, avatar, and transcription to facilitate communication with D/HH individuals. They decided to build on Amazon Web Services (AWS) for three reasons: the ability to scale up and down with customer demand, rapid compute for their artificial intelligence (AI) and ML needs, and cost-effective solutions for their substantial data corpus. “Our tech is hosted on AWS because we find the speed fantastic, as well as the compute,” Nicholas explains. “Primarily, we really like using AWS Trainium and AWS Inferentia because those are fantastic for low-cost training and inference. Additionally, they provide some of the fastest training and inference on the market, and are cost effective at that. We also find the infrastructure tools extremely helpful. Using Amazon Elastic Container Service (Amazon ECS) and AWS Fargate means that we can focus on just writing code, and AWS Code Pipeline ensures that we don’t need to worry about jumping through an arduous deployment process.”

The result? A solution that leads to easier communication between organizations and their customers. With Sign-Speak, businesses provide a better customer experience, improve their brand image, and comply with accessibility regulations.

Co-founders Yami and Nikolas demonstrate how Sign-Speak translates spoken words into ASL and vice versa.

Building an AI solution with (not for) the Deaf and Hard of Hearing community

Sign-Speak believes the power of AI comes from its ability to help their customers. They use ML to translate the visual language of ASL into spoken languages such as English. Additionally, with generative AI, they’ve created an avatar that translates spoken languages such as English into ASL. “I hope that Sign-Speak shows that AI technology can help a lot of communities that haven’t had access,” says Yami. “And at the same time, it will also show that helping these communities is a real business—it’s profitable, while also being good.”

Training Sign-Speak’s ML models within the Deaf community has been crucial to the development of the product. “Getting real-time feedback from the community is something we’re really proud of,” explains Yami. As their community engagement and product grew, they began accumulating a large corpus of data—much of which is videos—gathered from product pilots and through gamification. They recognized a growing need for powerful and scalable compute to support the features in their product roadmap.

With these challenges in mind, Sign-Speak applied for and was accepted into the AWS Impact Accelerator Latino Cohort, a program that provides under-represented founders with equitable access to funds, training, mentorship, tools, and resources.

Accelerating success with the AWS Impact Accelerator Latino Cohort

The Impact Accelerator was the on-ramp to expertise and support that would help Yami and her co-founders to prove both the profitability and the importance of technology such as Sign-Speak.

“We were very excited when AWS offered the opportunity to join the Impact Accelerator. It was a pivotal moment because of the business, technical, and monetary support from AWS,” Yami explains. “These helped us to solve a widespread problem: one that ranges from interpreter shortages to giving D/HH people access at places like restaurants and doctors’ offices.”

Going into the program, Sign-Speak had three goals:

Technical support. “It was very attractive to get a technical mentor for Nicholas, because a lot of accelerators don’t have that,” says Yami.
Networking and business support. Yami explains, “Making connections within an organization as big as Amazon has amazing implications for my business. I wanted to learn how an enterprise does things and apply that to my business as well.”
Financial support. “The cash grant and credits were critical for us to keep building,” says Yami. “And it was extremely attractive that the Impact Accelerator takes no equity from the startups they support.”

The eight-week program was “like nothing I’ve ever seen before, and I’ve been through a lot of accelerators,” says Yami. The intentionality of every piece of the program is what most stood out: from how mentors were assigned, to the one-to-one feedback that participants received about pitching and using Amazonian processes to improve their workflows.

“Every workshop with the Impact Accelerator was perfectly thought out,” she explains. “I met all three of my goals, and I also came out with a plethora of practical knowledge that allowed us to refine ourselves into better leaders and a better company.”

Yami’s proudest moment from the program is using what she learned to land a product pilot for Sign-Speak within a powerful company in her industry. “To see the drastic change in how we presented—from the start of the program to the end of the program—was amazing. And now I carry that skillset around with me and I use it.”

Gaining technical and business skills was important, and so was the atmosphere in which the cohort came together to grow as founders. “I think I cried four times in the cohort, because we formed such a strong relationship with each other,” says Yami. “Being around people who shared the same culture and faced similar challenges as me took it to the next level. It felt like a team and like a home.”

Yami attributes her Impact Accelerator experience to the mindset of AWS as a whole: “All Amazon thinks about is customers. And they got me into that mindset, too,” she laughs. “They understand fast iteration cycles to provide something of value—just like a startup does—and I think that’s what made this such a successful experience.”

Envisioning the future of Sign-Speak and Latino-led startups

Today, Sign-Speak continues to enhance their product in collaboration with the D/HH community. Yami believes the future of the interpretation industry will include the ability to have conversations anywhere at any time, for everyone: “AI has been a topic in communities of privilege for a long time. I want to change that—one day, I hope to see our AI technology built into enterprise products to enhance communication for all communities.”

For other founders who want to grow their startup with the help of an accelerator, Yami has three pieces of advice:

“Don’t build your product to make it a fit for the accelerator criteria. Instead, build it with your customer in mind, and everything else will follow.”
“Apply to accelerators that don’t take equity. It’s important that if you believe in yourself and your startup that you accept deals intentionally, and not because they’re easy or tempting.”
“Give constructive feedback to any programs you join. Make it better for the next stage of founders who participate. Founders are a community and we need to help each other.”

When it comes to supporting startup success, Yami advises: “I hope more companies take the leadership that AWS has taken to foster innovation and see the potential in people for greatness. When you’re dealing with communities such as the Latino community or the Black community, we’re people first: Developing a personal connection to us is great for business and it’s great for the community.”

For other Latino founders who are proving what’s possible with technology, Yami has a message:

“Pa’lante! We are innovators and we have grit. It’s hard, and funding is not always there, but I hope whoever is reading this understands: You can do it. It’s gonna be hard, but you can do it.”

Ready to begin your startup journey? Join AWS Activate to build and scale your startup with the right resources at the right time.

How STIGMA scaled their hope delivery app with the AWS Impact Accelerator

Megan Crowley — Mon, 18 Sep 2023 17:01:06 +0000

Many people think of hope as a synonym to optimism. The team at STIGMA—an asynchronous messaging app—embraces a different definition. Hope is not a feeling. It’s the act of acknowledging reality, even the really hard parts, and choosing to believe that your actions can make things better for you or for others. If hope is a skill you must practice, where can you practice it? That’s where the STIGMA app comes in.

Ariana Vargas, founder and CEO of STIGMA

STIGMA connects people who are struggling with strangers who share their lived experience to give them living proof that what they’re going through is something they will get through. While it’s estimated that 58 million – one in five adults – will be diagnosed with mental illness, the number of people who will struggle or feel overwhelmed and never reach the level of diagnosis is exponentially greater. STIGMA is proving what’s possible by transforming how and where people talk about mental health. Founded in 2021, STIGMA enables its thousands of members to share written, audio, and visual messages about mental health with each other within a moderated community. These conversations build trust, build empathy, and build hope.

“Our perception of the world is based on the lived experience we have and the stories we consume,” explains STIGMA founder and chief executive officer (CEO) Ariana Vargas, alluding to The Danger of a Single Story. She founded STIGMA to bridge the gap between mental health narratives and the lived experiences of the people who struggle with them.

“The idea behind STIGMA is ‘How do we connect people who have been through something hard with people who have also been through it to share their stories?’” says Ariana. “Especially when those people have done some work and made some progress and can help others do the same.”

As their user base continues to grow, STIGMA is leveraging lessons from their participation in the AWS Impact Accelerator Latino Founders cohort to scale their ability to share messages of hope by incorporating artificial intelligence (AI) and machine learning (ML) into their platform.

In addition to Crowdsourcing Hope® among their community, the STIGMA app works with vetted mental health and wellness providers to give members access to trusted mental health resources. “Making people feel seen and less alone is the purpose of the STIGMA app,” explains Ariana.

Creating a safe space for mental health storytelling

Becoming a founder was not part of Ariana’s life plans. Born in Costa Rica to an American mother and a Costa Rican father who suffered from schizophrenia, Ariana is one of the hundreds of millions of people who both live with mental illness, as well as love a person that lives with mental illness. Having experienced firsthand the power of empathy (and the negative effects of fear) on an individual’s wellbeing, Ariana’s initial goal was to pursue a medical degree. “I realized I loved helping people,” she explains.

Her studies at Boston College quickly pivoted from the scientific to the liberal arts. She leaned into her talent for storytelling, pursuing a film major and history minor. At the same time, Ariana’s serendipitous roommate assignment during freshman year became a best friendship that grew into a sisterhood.

While the two women were establishing their post-college lives, Ariana’s best friend was diagnosed with a rare and aggressive form of cancer. “Six months later, we had a conversation during which she asked me to promise I wouldn’t wait to live the life I wanted to live,” says Ariana.

Feeling compelled to take a step in the direction of the life she wanted, Ariana began planning a docuseries called “STIGMA,” where each episode showed a complete picture of someone living with a mental health condition. When the first short film premiered, featuring a survivor of sexual abuse who lives with complex post-traumatic stress disorder (PTSD) and depression, Ariana quickly saw that she’d achieved her goal of contributing meaningfully to the public conversations about mental health.

The STIGMA app

People were reaching out across social media and email to ask to speak to the film’s subject, as well as to share their own story. One man, in his sixties, was moved enough by the film to send a message to the film’s subject: “I can’t believe you talked about it. No one ever talks about it. The same thing happened to me and I never told anyone until now.”

“That’s the moment I became a startup founder,” Ariana explains. “Two years later, and every single time I tell that story the hairs on my arms raise. If that man had nowhere else to go in his life for all these years except for a private message to a stranger, then we are failing people who are struggling.”

Ariana left her full-time job in May of 2021 to begin working on STIGMA full time. She met STIGMA’s chief technology officer (CTO), Bill Guschwan, that summer and the two bonded over a vision for what healthy human connection that was facilitated by technology could look like. As a non-technical founder, she brought the business expertise and brand for STIGMA to life, while while a team of 12 contractors worked on creating an API-based application with features to build a safe and inclusive community for members:

Asynchronous communication to allow members to engage at their own pace
A moderated platform with no friending or private messaging to help members maintain healthy boundaries
Multi-modal content capacity (verbal, audio, and visual) to make the app accessible and build an inclusive community

Building and scaling with the AWS Impact Accelerator Latino Founders cohort

As the STIGMA team continued to integrate new features into their app, they were accepted into the AWS Impact Accelerator cohort for Latino founders.

STIGMA shared their goals with their two mentors—one for technology and one for business—to come up with solutions that would improve the product experience for their members while also helping them to hone skills such delivering a compelling pitch to investors, or the Amazon process for working backwards.

“What we received from the AWS Impact Accelerator is thought partnership in the two areas that are most important for a tech company at our stage: technology and business,” says Ariana. “It was an incredible opportunity to be plugged into an organization of AWS’ breadth of experience, working with some of the brightest minds in tech.”

One of STIGMA’s technical goals for the Impact Accelerator was to incorporate AI into their platform. The team envisioned a feature where members can receive immediate messages of hope that are labeled as AI-generated, while waiting for another member to answer with a human-generated response. “We collaborated with our technical lead and he sought answers from AI experts at AWS,” explains Ariana. Together, STIGMA and AWS worked to understand how large language models (LLMs) on their existing data sets of human-generated messages of hope could safely accelerate receiving a message of hope. Due to the sensitive nature of their platform, all computer generated messages are reviewed by human-in-the-loop moderators before being published. At the end of the eight-week accelerator, the STIGMA team launched this feature for their members, exponentially allowing them to help more people to feel seen and heard, when they need it most.

Ariana pitches during the AWS Impact Accelerator Investor Day to an audience of venture capitalists, supporters, and peers

At the end of the eight-week accelerator, Ariana and other members of the cohort pitched to a Demo Day audience that included investors, supporters, and peers. “I walked out of the program with the best pitch I’ve ever had, and I have the video to use again,” explains Ariana. “There are investors who say to me, ‘I saw the AWS pitch, let’s talk.’”

Accelerating the future of mental health with the STIGMA app

Ariana and the STIGMA team

After graduating from the AWS Impact Accelerator, STIGMA continued adding new features to their app to build their mental health community.

One feature currently in beta mode of the STIGMA app is “Narrate”, a narrative therapy inspired experience which will enable people to reframe the negative stories they’re telling about themselves into positive ones. The STIGMA team is implementing knowledge gained from their time in the AWS Impact Accelerator to better build this feature. During beta testing, 94% of people said that they experienced a meaningful increase in hope. “Our next step is to use those transcripts to understand how we can train an AI model to have those conversations, while keeping a clinician involved,” says Ariana.

Alongside building more product features, STIGMA will continue to advocate for education, compassion, and action through storytelling and trusted mental health resources. “We will always try to sit in that spot of being a lighthouse and a bastion of safety where people can come to listen or watch, even if they’re not ready to talk,” says Ariana. “STIGMA is a space to build compassion for what you know others are experiencing, or what you yourself are experiencing.”

“When I think about the future of the industry for mental health, I think the reduction of stigma is mandatory,” says Ariana. “We have to make people feel safe talking about what they’re going through.”

Ready to begin your startup journey? Join AWS Activate to build and scale your startup with the right resources at the right time.

Arcadia accelerates climate tech innovation while building on AWS

Megan Crowley — Fri, 01 Sep 2023 13:24:01 +0000

Data startups know that siloed data can lead to limited visibility that stifles innovation and collaboration, while increasing costs. Unified data allows organizations to derive insights, innovate quickly on behalf of customers, and operate more efficiently.

Kiran Bhatraju, Arcadia founder and CEO

Did you know that a lack of unified data is an issue that occurs both within companies as well as across industries? Arcadia, a climate technology company founded by Kiran Bhatraju in 2014, unifies energy utility consumption and pricing data, making the data accessible and useful to businesses and individuals.

Based on his experience at multiple clean energy startups—as well as his time spent shaping energy policy in Washington, DC—Kiran had seen firsthand how difficult it was to innovate in the clean energy space without a single source of utility data.

“Making sure that society moves away from fossil fuels is a huge problem,” Kiran explains. “I founded Arcadia because I did not see a way to transition to clean energy fast enough if utility data was not made available for all of the new clean energy companies coming to market.”

The concept of using software and data as tools in the fight against climate change is what inspired Anil Beniwal to join Kiran as Arcadia’s chief technology officer (CTO). “We have an opportunity to create a foundational data platform that can support the entire climate tech industry,” says Anil.

This climate tech unicorn makes the data easily accessible via their Arc platform and suite of API products. To democratize this access to energy data, Arcadia builds on Amazon Web Services (AWS) as their primary cloud provider.

In addition to bringing energy data to businesses, Arcadia also makes it easier for individuals and companies to join local solar farms—known as “community solar”—to use clean energy while also benefiting from solar credits on their utility bill.

Driving change with data

With thousands of utility providers in over 50 countries—and data coverage for 95% of US of residential and commercial utility accounts—Arcadia is proving that it’s possible to simplify the consolidation and impact of energy data.

Anil Beniwal, Arcadia CTO

Arcadia uses the data to make informed business decisions that support their company’s growth, to create impact in the clean energy industry, and to add value for their customers.

“Data is the heartbeat of our organization,” says Anil. With AWS, “We build sophisticated data acquisition, normalization, auditing, and delivery pipelines. These enable our customers to understand their own energy footprint, to provide new energy experiences to the end customers, and anything in between.”

The impact is substantial. Climate tech innovators are using Arcadia’s platform and API to:

Surface the best clean energy prices for commercial builders based on rates and tariffs
Create ESG (environmental, social, and governance) scores to help investors make more sustainable financial decisions
Help electric vehicle owners find the best-priced vehicle charging stations
Enable accurate prediction of solar panel production and costs

“All new energy products need to run through the “big grid”—the trillion-dollar infrastructure of poles and wires—which is probably one of the most incredible engineering feats in the country,” laughs Kiran. “Yet the data from it has never been available in a single source of truth. Arcadia is taking that pain point and making it into something simple for the first time.”

Accelerating Arcadia’s success with the help of AWS

With more than seven years of building on AWS, and as long-time members of AWS Activate—a program for startups— Arcadia agrees firsthand that AWS cloud is the right choice for climate tech companies to build, scale, and go to market.

Benefits such as AWS credits supported their work to build and scale in the early days, while an AWS Well-Architected review empowered them with key insights that helped them select the right AWS services to build up the micro-services stack they would best support their technical goals.

Building their tech stack

With all of their workloads running on AWS, Arcadia balances their technical goals with providing their customers with a great product experience.

“AWS allows us to scale seamlessly. We use a large swath of AWS services to quickly deploy, manage, and monitor our applications,” explains Anil. “Being able to quickly deploy new services and new capabilities—with the push of a button—and scale them to meet demands means we can be more responsive to the needs of our customers and our business.”

On the compute side, Arcadia relies heavily on Amazon Elastic Kubernetes Service (Amazon EKS) for their Kubernetes workloads. Their data acquisition workflows also rely on AWS Lambda, a serverless event-driven service that allows you to run code without managing or provisioning servers.

For queuing, Arcadia uses Amazon Simple Queue Service (Amazon SQS). They chose Amazon EventBridge, a serverless event router, for event streaming between their services.

To meet their database and storage needs, Arcadia primarily uses Amazon Aurora, along with Amazon DynamoDB. Arcadia keeps their historical data and files stored in Amazon Simple Storage Service (Amazon S3) buckets.

Machine learning is a tool that allows Arcadia to enhance positive outcomes for their customers. “When we take all of that data, we use tools like Amazon SageMaker to help us drive more customer value,” says Anil. “We also use SageMaker to add more efficiency for our internal team.”

Arcadia also leverages many of the AWS security services to ensure a security posture that is compliant with their internal policies, as well as industry best practices.

Optimizing the cost of the cloud

As is the case with many startups, saving money while optimizing their product is a priority for Arcadia. Throughout their company journey, Arcadia has kept a consistent focus on right-sizing their AWS infrastructure to support both their business priorities and customer needs. Additionally, they use AWS tools such as Savings Plans and Amazon EC2 Reserved Instances to improve efficiency and get the most out of their cloud spend.

“By working with the AWS team, we’ve identified ways to bring our monthly costs down through things such as architectural reviews, reserved instance purchases, and savings plans,” Anil explains. “With the AWS team, we were able to find some real opportunities to drive down our costs without sacrificing at all on our customer value.”

Planning for their future growth

Along with building their tech stack on AWS, Arcadia works with their AWS account team to find opportunities that can grow their customer base and provide a better customer experience.

“The startups organization within AWS has been helping us to find new opportunities to scale with our customers,” explains Anil. “One way is by leaning into an offering on the AWS Marketplace as another procurement channel where customers can sign up for the Arc platform.”

Inspiring replicable success

What tenets help Arcadia to succeed in the clean energy sector? Kiran explains there are three important table stakes to building a successful and long-term company:

“Great companies are built by great people.” Having spent a decade leading Arcadia to its current size, Kiran recognizes the importance of “making sure the talent bar is consistently high for your company in its earliest stages, and definitely in its later stages.”
A strong and aligned relationship with your startup’s investors is equally important. “Choose your investors wisely and make sure they’re aligned with both the mission and what you want to build,” Kiran explains.
For early-stage founders especially, Kiran advises that “You have to really care about what you’re building or else you’re going to lose. Be committed to getting through the tough times and be in it for the long haul.”

Building a greener future

Market opportunities for climate tech startups continue to grow as large companies and consumers take a stronger stance on the importance of sustainable choices. By 2027, the climate tech market is expected to reach nearly $1.4 trillion, representing a compound annual growth rate of 8.8%.

Kiran cites the Inflation Reduction Act, which passed in August of 2022, as an example. “That is the largest climate investment by any country in history, with a financial impact between $350 million and $1 trillion,” he says. “It’s a hard opportunity for startups to ignore.”

As others are starting to build, Arcadia plans to continue growing. “One of the things we’re really focused on is making sure that customers have access to all of the data they need,” Kiran says. This includes growing Arcadia’s community solar market, continuing to make their APIs developer-friendly and easy to use, and exploring how working with AWS to incorporate new technology such as artificial intelligence can help Arcadia bring more to the table for its customers.

“Climate tech is the most exciting place to build a career right now,” says Kiran. “I think decarbonization will be bigger than the internet because it will touch every single sector and piece of the economy.”

Ready to begin your startup journey? Join AWS Activate to build and scale your startup with the right resources at the right time.

Learn more about how startups are using data solutions on AWS:

Building generative AI applications for your startup, part 2

Hrushikesh Gangur — Thu, 17 Aug 2023 11:56:11 +0000

This blog series in two parts discusses how to build artificial intelligence (AI) systems that can generate new content. The first part gives an introduction, explains various approaches to build generative AI applications, and reviews their key components. The second part maps these components with the right AWS services, which can help startups quickly develop and launch generative AI products or solutions by avoiding time and money spent on undifferentiated heavy lifting work.

Let’s advance this blog series, and dive deep into Amazon Web Services (AWS) capabilities your startup can leverage to get your product to market as quickly as possible, while keeping cost efficiency and performance as key goals.

Which AWS services should I use to build my generative AI application?

This can be well explained through my illustration of generative AI components from part 1 of this blog series. The following diagram, Figure 1, maps each component to corresponding AWS service(s). Note that these are a curated set of AWS services I see startups taking benefits from; however, there are other AWS services available.

Figure 1: Mapping AWS services to generative AI components.

To elaborate, I will start with mapping AWS services to the common components of a generative AI application. Then I will explain the AWS services that map to the remaining components in Figure 1, based on the approaches you use to implement your application.

Common components

The common components of a generative AI application are the foundation model (FM), its interface, and optionally the machine learning (ML) platform and accelerated computing. These can be met using managed offerings available from AWS:

Amazon Bedrock (foundation model and its interface components)

Amazon Bedrock, a fully managed service that makes foundation models from leading AI startups (AI21’s Jurassic, Anthropic’s Claude, Cohere’s Command and Embedding, Stability’s SDXL models) and Amazon (Titan Text and Embeddings models) available via API, so you can choose from a wide range of FMs to find the model that’s best suited for your use case. Amazon Bedrock provides API or serverless access to a set of foundation models to provide three capabilities: text embedding, prompt/response, and fine-tuning (on select models).

Figure 2: Amazon Bedrock workflow

Amazon Bedrock is well-suited for application or model consumer startups who are building value-added services – prompt engineering, retrieval-augmented generation, and more – around a foundation model of their choice. Its pricing model is pay-by-use, typically in the unit of millions of tokens processed. Amazon Bedrock is generally available; however some of the features discussed in this blog are in private preview. Learn more here.

Amazon SageMaker JumpStart (foundation model and its interface components)

AWS offers generative AI capabilities to Amazon SageMaker Jumpstart: a foundation model hub containing both publicly available and proprietary models, quick start solutions, and example notebooks to deploy and fine-tune models. When you deploy these models, it creates a real-time inference endpoint which you can access as directly using SageMaker SDK/API. Or, you can front-end SageMaker’s foundation model endpoint with AWS API Gateway and a lightweight compute logic in an AWS Lambda function. You can also leverage some of these models for text embedding.

Figure 3: Amazon SageMaker JumpStart workflow

Both the inference endpoint and the fine-tuning training jobs run on your choice of managed ML instances (see “Accelerated Computing” in Figure 1) using SageMaker as the ML platform (see “ML Platform” in Figure 1). SageMaker Jumpstart is well-suited for application or model consumer startups who want more control over their infrastructure, and who have moderate ML skills and infrastructure knowledge. Its pricing model is pay-by-use, typically in the unit of instance-hours. All the models and solutions in this offering are generally available.

Amazon SageMaker training and inference (ML platform)

Startups can leverage Amazon SageMaker’s training and inference features for advanced capabilities like distributed training, distributed inference, multi-model endpoints, and more. You can bring the foundation models from the model hub of your choice – whether that’s SageMaker JumpStart or Hugging Face or AWS Marketplace, or you can build your own foundation model from scratch.

Figure 4: Amazon SageMaker training and inference workflow

SageMaker is well-suited for full-stack generative AI application builders (from model providers to model consumers), or for model providers with teams who have advanced ML and data pre-processing skills. SageMaker also offers a pay-by-use pricing model, typically in the unit of instance-hours.

AWS Trainium and AWS Inferentia (accelerated computing)

In April 2023, AWS announced general availability of Amazon EC2 Trn1n Instances powered by AWS Trainium, and Amazon EC2 Inf2 Instances powered by AWS Inferentia2. You can leverage AWS purpose-built accelerators (AWS Trainium and AWS Inferentia) using SageMaker as the ML platform.

The benchmark testing for inference workloads reports Inf2 instances perform with 52% lower costs against a comparable inference-optimized Amazon EC2 instance. I suggest keeping an eye on fast development cycles of AWS Neuron SDK, where approximately every month AWS is adding new model architecture in their support matrix for both training and inference.

Approaches for building generative AI applications

Now, let’s discuss each of the components in Figure 1 from an implementation perspective.

The zero-shot or few-shot learning inference approach

As we discuss in part 1, zero-shot or few-shot learning is the simplest approach for building a generative AI application. To build applications based on this approach, all you need are the services for the four common components (foundation model, its interface, ML platform, and compute), your custom code to generate prompts, and a front-end web/mobile app.

Figure 5: Components of the zero-shot learning approach

To learn more about selecting a foundation model through Amazon Bedrock or Amazon SageMaker JumpStart, refer to the model selection guidelines here.

The custom code can leverage developer tools like LangChain for prompt templates and generation. The LangChain community has already added support for Amazon Bedrock, Amazon API Gateway, and SageMaker endpoints. Just to remind you, you may also like to leverage AWS Amazon CodeWhisperer, a coding companion tool, to help improve developers’ efficiency.

Startups building a front-end web app or mobile app can easily start and scale by using AWS Amplify, and host these web apps in a fast, secure, and reliable way using AWS Amplify Hosting.

Check out this example of zero-shot learning that builds with SageMaker Jumpstart.

The information retrieval approach

As discussed in part 1, one of the ways your startup can customize foundation models is through augmenting with an information retrieval system, most commonly known as retrieval-augmented generation (RAG). This approach involves all of the components mentioned in zero-shot and few-shot learning, as well as the text embeddings endpoint and vector database.

Figure 6: Components of the information retrieval approach

Options for the text embeddings endpoint vary depending on which AWS managed service you’ve selected:

Amazon Bedrock offers an embeddings large language model (LLM) that translates text inputs (words, phrases, or possibly large units of text) into numerical representations (known as embeddings) that contain the semantic meaning of the text.
If using SageMaker JumpStart, you can host an embeddings model like GPT-J 6B or any other LLM of your choice from the model hub. The SageMaker endpoint can be invoked by the SageMaker SDK or Boto3 to translate text inputs into embeddings.

The embeddings can then be stored in a vector datastore to do semantic searches using either Amazon RDS for PostgreSQL’s pgvector extension, or Amazon OpenSearch Service’s k-NN plugin. Startups prefer one or the other based on which service they are typically most comfortable using. In some cases, startups use AI native vector databases from AWS partners or from open source. For guidance on vector datastore selection, I recommend referring to The role of vector datastores in generative AI applications.

In this approach too, developer tools play a pivotal role. They provide an easy plug-n-play framework, prompt templates, and wide-support for integrations.

Going forward, you can also leverage agents for Amazon Bedrock, a new capability for developers that can manage API calls to your company systems.

Check out this example of using retrieval augmented generation with foundation models in Amazon SageMaker Jumpstart.

The fine-tuning or further pre-training approach

Now, let’s map the components to the AWS services needed for the last approach to implementing a generative AI application: fine-tuning or further pre-training a foundation model. This approach involves all of the components discussed in zero-shot or few-shot learning, as well as data pre-processing and model training.

Figure 7: Components of the fine-tuning or further pre-training approach

Data preparation (sometimes called preprocessing or annotation) is particularly important during fine-tuning, where you need smaller and labeled data sets. Startups can easily get started using Amazon SageMaker Data Wrangler. This service helps reduce the time it takes to aggregate and prepare tabular and image data for machine learning from weeks to minutes. You may also leverage this service’s inference pipeline feature to chain the preprocessing workflow to training or fine-tuning jobs.

If your startup needs to preprocess a huge corpus of unstructured and unlabeled datasets in your data lake on Amazon S3, you have a few options:

If you’re using Python and popular Python libraries, is useful to leverage AWS Glue for Ray. AWS Glue uses Ray, an open source unified compute framework used to scale Python workloads
Alternatively, Amazon EMR can help process vast amounts of data using open source tools such as Apache Spark, Apache Hive, Apache HBase, Apache Flink, Apache Hudi, and Presto.

For the model training component of this approach, Amazon Bedrock allows you to privately customize FMs with your own data. It manages your FMs at scale without having to manage any infrastructure (this is the API way to fine-tune). Alternatively, the SageMaker Jumpstart approach provides a quick-start solution to privately fine-tune (on select models) for instruction or domain adaptation using your own data. You can modify the SageMaker JumpStart bundled training script for your needs, or you can bring your own training scripts for open-source models, and submit these as SageMaker’s training job. If you have to further pre-train the model (typically for open source models), you can leverage SageMaker’s distributed training libraries to speed up and efficiently utilize all of the GPUs of an ML instance.

In addition, you may also consider fully managed data generation, data annotation services, and model development with the Reinforced Learning from Human Feedback technique using Amazon SageMaker Ground Truth Plus.

An example architecture

So, how do all of these components look when realizing a generative AI use case? While every startup has a different use case, and unique approaches to solving real world problems, one common theme or starting point I have seen in building generative AI applications is the retrieval-augmented generation approach. After plugging in all those AWS services discussed above, the architecture looks like this:

Ingestion pipeline – The domain-specific or proprietary data is preprocessed as text data. It is either batch processed (stored in Amazon S3) or streamed (using Amazon Kinesis) as it is created or updated through the embedding process, and stored in dense vector representation.

Figure 8: An example ingestion pipeline for a generative AI application.

Retrieval pipeline – When a user queries the proprietary data stored in vector representation, it retrieves the related documents using k nearest neighbor (kNN) or semantic search. It is then decoded back to clear text. The output serves as rich and dense context to the prompt.

Figure 9: An example retrieval pipeline for a generative AI application.

Summarization generation pipeline – The context is added to the prompt with the original user query to get insight or summarization from the retrieved document.

Figure 10: An example summarization generation pipeline for a generative AI application.

All of these layers can be built with a few lines of code by using developer tools like LangChain.

Conclusion

This is one way to build an end-to-end generative AI application using AWS services. The AWS services you select will vary based on the use case or customization approach you take. Stay tuned on latest AWS releases, solutions, and blogs in generative AI by bookmarking this link.

Let’s go build generative AI applications on AWS! Kickstart your generative AI journey with AWS Activate, a free program specifically designed for startups and early stage entrepreneurs that offers the resources needed to get started on AWS.

Compute for Climate Fellowship to fund new tech solutions addressing the climate crisis

Lisbeth Kaufman — Thu, 03 Aug 2023 20:42:54 +0000

The program will select and fully fund proof of concepts for new ideas that leverage advanced cloud computing, including generative AI and complex systems modeling and simulation, to solve some of the biggest challenges in the fight against climate change.

The United Nations (UN) Secretary-General has called climate change “the defining issue of our time.” We couldn’t agree more. Earlier this month, the world recorded the highest average temperatures across the globe, according to data from the U.S. National Centers on Environmental Prediction. Wildfires, floods, droughts, and rising sea levels have been consistently ravaging different parts of the planet for years now, which poses a real threat to food security, people’s health, and biodiversity.

Solving these challenges requires innovation. We believe startups are well suited to help find new solutions to address the climate crisis. The venture capital community agrees: according to PWC, 25% of all venture capital investments in 2022 were directed toward climate tech startups.

While thousands of these innovators are inventing new science and technologies every day to address this urgent issue, the research and development (R&D) needed to bring them to market can be an expensive and lengthy process. To help accelerate some of the brightest ideas out there, Amazon Web Services (AWS) is partnering with the International Research Centre on Artificial Intelligence (IRCAI), an organization under the auspices of United Nations Educational, Scientific and Cultural Organization (UNESCO), to launch the Compute for Climate Fellowship. This first-of-its-kind global program is designed to support entrepreneurs and startups that are applying advanced cloud computing and artificial intelligence (AI) to create new solutions that address the climate crisis.

The Compute for Climate Fellowship select proposals that think big, have the most innovative use of advanced cloud computing, and the biggest promise for global impact within six key problem areas in the fight against climate change:

Climate risk and resilience
Food security
Biodiversity and conservation
Climate-related health
Circular economy
Environmental, social, and governance (ESG) analytics

Through the fellowship, IRCAI and AWS will provide global climate tech startups access to various technical resources to build their POCs, including:

A team of AI, sustainability, and ethics experts
Access to advanced computing services, such as high-performance computing (HPC)
Cloud computing products and services that support AI, generative AI, and machine learning (ML) solutions
AWS Activate credits to cover the POC build

In addition, all POCs will be designed under the guidelines of UNESCO’s Ethics Impact Assessment for AI. This ensures that each solution is built with safe, trustworthy technology.

“With this program, we want to give those founders and innovators trying to solve some of the most complex and pervasive challenges, varying from materials science to agriculture and weather analytics to biodiversity remediation, a leg up in bringing their ideas to life,” said Kathryn Van Nuys, Global Head of Specialized Segment Startups at AWS. “Beyond funds to cover their POCs and technical support, by partnering with IRCAI, we also ensure that they have access to best practices and know-how on bringing ethical and safe products and services to market.”

Advanced computing and AI have already shown great promise in helping create solutions to fight climate change. AWS is currently supporting customers in developing early-warning systems for food insecurity; analytics for health equity in the context of climate change; intelligence to power the circular economy; and modeling and simulation capabilities that use HPC to identify and mitigate climate-related fires, floods, and heat waves.

AWS is also helping customers explore how large language models can help remediate biodiversity loss, support mental health interventions in impacted populations, and foster innovation among automotive manufacturers focused on reaching net-zero goals, electrification, and circularity.

One example is Good Chemistry. The startup discovered new pathways to destroy per- and polyfluoroalkyl substances (PFAS). These “forever chemicals,” don’t break down naturally and can cause harm to people’s health and the environment. Using AWS quantum and HPC computing, Good Chemistry has proved that “forever chemicals” don’t actually need to last forever. With this technology, they can run simulations using more than a million CPU cores to calculate energies for bond-breaking in several PFAS.

BlocPower analyzes over 100 million buildings’ energy profiles to better understand how to optimize energy efficiency across the United States. They source energy profiles using EnergyPlus, the US Department of Energy’s open-source whole-building modeling engine.

With AWS HPC capabilities, BlocPower has processed over 30 TB of data at speeds 16,000 times greater than they could before. Within three months, the startup finished building its entire data pipeline in the cloud. This empowered them to deploy BlocMaps, a SaaS solution that provides actionable insights for property owners, utility companies, municipalities, states, and other groups interested in decarbonizing buildings.

Applications are open now

Applications for the Compute for Climate Fellowship are open now! Submissions will be accepted and selected on a rolling basis. Submit your application before August 31 to be considered and maybe have the opportunity to be showcased at AWS re:Invent conference in November 2023. Applications submitted after September 1 will be considered for development in 2024.

Startups that apply but are not selected to participate will have access to up to $5,000 in AWS credits through the AWS Activate program, as well as free workshops and training on how to use AWS HPC services to build their solutions.

To apply for the Compute for Climate Fellowship and learn details on eligibility, visit https://ircai.org/compute-for-climate-fellowship.

Building a serverless dynamic DNS system with AWS

Zac Burns — Wed, 02 Aug 2023 20:18:10 +0000

This blog post was originally published in December 2015. It was updated July 2023 to make the solution more cost-effective and efficient. This post has been updated to replace Amazon API Gateway with AWS Lambda function URLs and Amazon Simple Storage Service (Amazon S3) with Amazon DynamoDB. Using Lambda function URLs reduces the overall cost of the solution. This feature comes at no extra cost when using the Lambda service, and it provides a RESTful HTTPs endpoint that our client interacts with. Replacing Amazon S3 with DynamoDB increases the efficiency of the solution and reduces the overall latency when querying data.

Early-stage startups, small businesses, and home networks often have dynamic public IP addresses that can change without notice. Because of this changing address, you can’t reliably access systems on these networks from the outside. For startups in the early stages of their life cycle, it’s important to provide a reliable and highly available service in order to gain trust with your first set of customers.

Dynamic DNS systems solve this problem by running a software agent inside your network to keep a DNS record updated with the latest public IP address. As long as the DNS record is up to date, you can find your network and customers can reliably access your service.

In this post, we describe how to build your own dynamic DNS system by using serverless AWS services. Building a serverless system using nothing but Amazon Web Services (AWS) services and a few lines of code is simple, cost-effective, and scalable, and allows you to focus on the core business logic of your startup, rather than worrying about scaling and maintaining the underlying infrastructure.

AWS services we use in our dynamic DNS system

In the next sections, we show you how to use the following AWS services to build a dynamic DNS solution:

The AWS Lambda service allows you to run code without having to manage the underlying servers. Your code is always ready to run, but you are charged only per invocation of the function, in 1 millisecond increments. The Lambda service can interact with other AWS services through AWS SDKs.
Lambda function URLs provide a dedicated HTTPS endpoint for your Lambda function. This allows you to directly invoke the function from your client application without the need to use an AWS SDK or invoke the function via an additional proxy service. This feature comes at no additional cost to the Lambda service.
Amazon Route 53 is a managed DNS service that allows you to register and host domains and DNS zones from a global network of DNS servers. As with all AWS services, Route 53 can be managed through APIs.
DynamoDB is a fully managed, serverless, key-value NoSQL database designed to run high-performance applications at any scale. DynamoDB offers built-in security, continuous backups, automated multi-Region replication, in-memory caching, and data import and export tools.

Logical flow of the dynamic DNS system

Figure 1 shows how a client finds its own IP address by making an API request to a service built using a Lambda function and its attached function URL.

Figure 1. Request flow for client to retrieve IP address of its running service

As shown in Figure 2, now that the client knows its public IP, it makes another request to our service to set a DNS record. The Lambda function first consults the record stored in our DynamoDB table to validate the request. If the check passes, the Lambda function then sets the DNS entry in Route 53 via an API call. Now the network’s current IP is in public DNS and can be found by a standard DNS query.

Figure 2. Request flow to set update DNS record with new IP

Benefits of dynamic DNS with Lambda and Route 53

Here’s some advantages you will gain by running a serverless dynamic DNS system:

Easier to set up. There is a sample client along with all the code, configuration, and instructions to set this up in your own AWS account.
Thin to no client. It takes three commands to update the API. You can write your own client in most languages and run it on platforms, including Windows, Linux, macOS, Raspberry Pi, Chrome OS, and DD-WRT/Tomato USB router firmware.
Support for arbitrary number of clients, hostnames, and domains. Route 53 service limits and quotas can be found in the “Quotas” section of the Route 53 Developer Guide.
Cost-effective–$1 to $2 per month. Route 53 zones each cost $0.50 per month, 1,000,000 DNS queries cost $0.40, and 10,000 Lambda function requests to update DNS cost under $0.01.
Serverless architecture. Serverless technologies feature automatic scaling, built-in high availability, and a pay-for-use billing model to increase agility and optimize costs.
Granular permissions allow only authorized clients to update their own hostname. Clients can update the system only from the address that is being added to DNS.
Minor changes required for your current DNS setup. You can leave your primary com zone with your current DNS provider and use a secondary dynamic.example.com zone in AWS. Refer to “Creating a public hosted zone” in the Route 53 Developer Guide for more information on creating hosted zones in Route 53.

Prerequisites

You’ll need two things to build this solution:

An AWS account. New accounts are eligible for the AWS Free Tier.
A domain you own, hosted on Route 53 or another provider. You can register domains through Route 53 if needed from as little as $3.00. Refer to Amazon Route 53 Pricing for Domain Registration for a full range of prices.

How to build a dynamic DNS system in your account

At this point, you have enough information to start building your own copy of the system. If you want to learn more about how it works, read on.

If you want to start building, Route 53 dynamic DNS with Lambda is available on GitHub. It provides illustrated instructions and all the necessary code and configuration.

How the new system works

First, the client needs to find the public IP assigned to its network. If you make a request from your network to a service on the internet, that service sees the request coming from your external IP address.

Then, the client makes an HTTP POST request to the Lambda function URL with a request body of {"execution_mode":"get"}, and gets a response containing the current public IP address:

HTTP POST
> https://.....lambda-url.eu-west-1.on.aws

{“return_message”:
“176.32.100.36”, “return_status”: “success”}

During this process, the Lambda function URL converts the HTTP request to JSON, including all request parameters, and passes the requestor’s source IP address to a Python Lambda function. The Lambda function then sends a JSON response with the IP back to the client.

Figure 3 shows a request to get a public IP.

Figure 3. Retrieving public IP address

The client now builds a request token by linking the public IP address returned from the HTTP POST request, the DNS hostname, and a shared secret. For example, if your IP address is 176.32.100.36, your hostname is host1.dyn.example.com, and your shared secret is shared_secret_1, the linked string will be the following:

176.32.100.36host1.dyn.example.comshared_secret_1

Next, the client generates a SHA-256 hash function from the string:

echo -n 176.32.100.36host1.dyn.example.comshared_secret_1 | shasum -a 256

Hash: 96772404892f24ada64bbc4b92a0949b25ccc703270b1f6a51602a1059815535

The client then requests the DNS update by making a second HTTP POST request. It passes the plain text hostname as a key and the hash function as the authentication token within the request body:

HTTP POST > https://....lambda-url.eu-west-1.on.aws

{“execution_mode”:”set”, “ddns_hostname”:”host1.dyn.example.com”, “validation_hash”:”96772404892f24ada64bbc4b92a0949b25ccc703270b1f6a51602a1059815535”}

The Lambda function URL then passes the request parameters back to the Lambda function.

After that, the Lambda function queries its JSON configuration record from DynamoDB using the AWS SDK for Python (Boto3). In this system, interactions between the Lambda service, DynamoDB, and Route 53 use Boto3, which is pre-built into the Lambda service runtime environment.

Once our Lambda function queries the configuration record from DynamoDB, it uses the hostname as a key to find the shared secret, and other configuration associated with that record, similar to the following example record:

{
"host1.dyn.example.com.": {
"aws_region": "us-west-2",
"route_53_zone_id": "MY_ZONE_ID",
"route_53_record_ttl": 60,
"route_53_record_type": "A",
"shared_secret": "SHARED_SECRET_1"
},
"host2.dyn.example.com.": {.....

The client passed host1.dyn.example.com as the key, so the Lambda function reads SHARED_SECRET_1 from the configuration, and rebuilds the hash function token using the hostname, the requestor’s IP address, and the shared secret. If the hash function calculated by the Lambda function and the hash function received from the client match, then the request is considered valid.

Once the request is validated, the Lambda function uses the information from the configuration to make an API call to Route 53 to see if the DNS hostname is already set with the client IP. If no change is necessary, the Lambda function responds to the client and exits:

{“return_message”: “Your IP address matches the current Route53 DNS record.”, “return_status”: “success”}

If there is no record, or if the current record and the client IP do not match, the Lambda function makes an API call to Route 53 to set the record, responds to the client, and exits:

{“return_message”: “Your hostname record host1.dyn.example.com. has been set to 176.32.100.36”, “return_status”: “success”}

Figure 4 shows the request to set the hostname.

Figure 4. Request flow to set hostname

How is this system secured?

All communications using Lambda function URLs are encrypted.
The shared secret is never transmitted across the internet.
Client request rate can be throttled using the reserved concurrency feature of the Lambda service.
The authentication mechanism is multi-factor because the client presents the shared secret (“something it has”) and its own public IP address (“something it is”).
The configuration file can be encrypted at rest via DynamoDB server-side encryption.
Your AWS credentials are not used, so they cannot be leaked.

Conclusion

The dynamic DNS system we describe in this post shows how to create your own serverless solution on AWS to solve a real-world problem—DNS is susceptible to change and you might not be aware!

Use this solution to run your own dynamic DNS on AWS. Or, use it as an example to learn how you can use AWS services to create your own serverless solutions at any scale.

Visit the Route 53 dynamic DNS with Lambda repository on GitHub for a complete set of code, configuration, and instructions on how to deploy.

Dune builds on AWS to amplify the impact of blockchain data

Mats Olsen — Tue, 11 Jul 2023 12:42:33 +0000

Startups are often founded by fearless individuals who face a problem and—instead of being deterred—set out to solve it for themselves and for everyone else.

Mats Olsen, co-founder and CTO of Dune

Dune, a web3 analytics unicorn founded in 2018, is one such startup. While building Ethereum prototypes on the blockchain, co-founders Fredrik Haga and Mats Olsen recognized that disparate sources of crypto data are a major impediment to web3 development. Mats, who is also Dune’s chief technology officer (CTO), explains: “While building, one of our biggest pain points was getting structured information back from the blockchain. It’s a database that’s optimized for writing, but not for reading.”

Dune is proving it’s possible to simplify the consolidation and analysis of blockchain data. Dune builds on Amazon Web Services (AWS) to provide a web-based platform that allows people to query public blockchain data and aggregate it into shareable dashboards. Users can take cross-chain data (data from separate blockchains) for multiple tokens, wallets, and protocols and build a dashboard that makes the data more transparent and actionable.

In the beginning, Dune focused on taking crypto data and making it easily available and accessible to other web3 developers. As Mats explains, “Startups in the space had to duplicate a lot of work and engineering hours just to ask questions like, ‘How many users do I have?’” As Dune went to market with a product that made crypto data easier to understand and use, they began adding visualization and dashboarding elements to their platform to make the data even more actionable.

“In crypto, a lot of people say that the data is available and you can just look at it. It’s true that the data is available, but Dune makes it practical to use it,” says Mats. “I’m proud of the transparency and accessibility we’ve brought into this space.”

Today, Dune is known for dashboards created by community members that are shared virally across social media, crypto news websites, and on Dune itself. “One of the most interesting aspects of the crypto space is the viral aspect of it,” Mats says. “There’s a so-called stakeholder explosion happening where public blockchain usage data is valuable to the product owners, but also to its investors, competitors, and regulators. It’s an interesting phenomenon.”

An image of a Dune dashboard. Check out the interactive version at https://dune.com/oplabspbc/soundxyz-on-optimism

Building and scaling by migrating to AWS

As Dune’s user base expanded and their technical needs grew, Mats and team decided that it was time to migrate to an all-in-one cloud provider from their multi-cloud setup. Dune chose to go all-in on AWS.

One of Dune’s primary reasons for migrating was to host workloads in a single location with services that could meet both their present and future needs. Mats explains, “The fact that AWS has such a wide range of offerings was really important for us. AWS does services really well—like DynamoDB, for instance. We don’t use it a whole lot right now, but it’s in our future plans.”

Along with offering over 200 fully featured services—more than any other cloud provider—AWS is the most broadly adopted cloud. Like most startups, Dune wants their engineers to be able to innovate and build as quickly as possible.

“AWS is so well-known and so prolific in the engineering landscape. It’s easy because most of our engineers have used it before. There’s no ramp-up time and that is really helpful,” says Mats.

Dune leveraged the AWS Migration Acceleration Program (MAP) and AWS Activate to migrate to AWS. Credits helped Dune migrate to AWS with no costs in the first three months.

Today, Dune’s tech stack includes Amazon Simple Storage Service (Amazon S3)—object storage that allows any amount of data to be retrieved from anywhere–to host the enormous amounts of data that contribute to their dashboards. To process this data, Dune uses Apache Spark and then queries the data with Trino. They also run a lot of Kubernetes clusters, which is why most of Dune’s applications use Amazon Elastic Kubernetes Service (Amazon EKS). To implement secure and scalable customer identity and access management, Dune uses Amazon Cognito.

By building on AWS, Dune gets support from the AWS Startups Team for everything from infrastructure and optimization, to go-to-market and networking. “We’ve always been very happy with the help we’ve gotten from AWS,” says Mats. “It’s nice to know the support is there for us, be that with the solutions architects or commercial opportunities.”

Focusing more on product and less on spend

Dune is all about the data, including when it comes to their cloud costs: “The biggest thing for us is understanding where we spend money and how much value that adds for our customers,” says Mats.

To reduce cloud costs and make the most of their spending, Dune worked with their account team to choose options that made the most financial sense for their technical needs. These include AWS Graviton processors and Karpenter, an open-source Kubernetes cluster autoscaler.

Additionally, Dune was able to save 26% on their Amazon EC2 Reserved Instances (when compared to on-demand usage). They accomplished this by automating the purchasing of these instances with Zesty and by using Savings Plans, which is a flexible pricing model for AWS compute usage. Mats also notes that, “Spot Instances have also been a big part of our strategy to reduce cloud costs because they help with our high Kubernetes use.”

AWS Activate—a program that offers startups free tools, resources, and more—are also a key to Dune’s ability to focus less on spend and more on product. “The credits were incredibly helpful for us to worry less about optimizing our cloud workloads,” says Mats. With the credits they received from AWS Activate, Dune was able to provide their customers with a better product and increase innovation.

“What the credits most importantly allowed us to do was focus on product market fit. To focus on improving our products for our users’ needs, instead of worrying about managing spend,” Mats explains. “Credits lowered the barrier to trying out and experimenting with new products because we didn’t immediately have a bill for them, so to speak.”

Lessons for founding a successful startup

As Dune showed with their use of AWS Activate credits, building products that solve a problem for your customer is rule number one. Mat explains, “The only thing that matters are your users. If you don’t have any, find some.” The best way to do this is to find a problem and solve it for yourself, as well as the people in that space. He advises that listening to your customers is equally important, as is balancing their feedback with the long-term vision that you have for your product.

Founding a successful startup is not an easy process. “Most startups die because you give up,” advises Mats. “You have to stay in the game if you want to win.” He recounts how in the early days, Dune’s team members worked for nearly a year without a salary, they had only three customers, and at one point the company was weeks away from closing their doors. They didn’t give up, though, and an investment came through that allowed Dune to keep building and growing.

BUIDLing the future of web3 together

Dune’s first five years were an exciting time. Alongside supporting the data needs during two major consumer waves in crypto–DeFi (decentralized finance) and NFTs (non-fungible tokens)–Dune simultaneously scaled from an idea to a unicorn in only three years. Now, says Mats, the time has come for technology to be “more important than ever” in contributing to web3 innovation, currencies, and how people store value.

Dune plans to rise to the occasion. With more than 40 employees, they have the capacity to incorporate new features to make crypto data even more accessible to their users. “The ability to scale compute on our backend to meet our users’ needs, is important,” says Mats. “We’re investing a lot in making autoscaling provide an even better customer experience.”

What else does the future hold? “What I’m most excited about is generative AI,” shares Mats. “In June, I began leading an initiative to experiment with how large language models can allow users to interact with our database without using SQL.” The ability to generate SQL queries from natural language text, known as text-to-SQL, will lower Dune’s barrier to entry: This will allow users who do not know how to write and run SQL to interact with their database.

Generative AI is a new lever that can help to democratize web3 by allowing more people than ever before to aggregate and visualize cross-chain data. “When we’re ready to experiment with open source large language models, hosting them and fine-tuning them on AWS makes the most sense for us,” says Mats.

Whether it’s optimizing their tech stack with the help of AWS, building new products, or innovating with technological advancements—Mats knows the Dune team is ready to tackle all opportunities the future brings.

“What sets us apart from a lot of other companies is our ability to take technology and modify it to our needs. We’re not afraid to use open source tooling and we’re happy to get deep into the weeds. All of this leads to a better user experience,” Mats says.

Building generative AI applications for your startup, part 1

Hrushikesh Gangur — Wed, 05 Jul 2023 13:52:03 +0000

This blog series in two parts discusses how to build artificial intelligence (AI) systems that can generate new content. The first part gives an introduction, explains various approaches to build generative AI applications, and reviews their key components. The second part maps these components with the right AWS services, which can help startups quickly develop and launch generative AI products or solutions by avoiding time and money spent on undifferentiated heavy lifting work.

Recent generative AI advancements are raising the bar on tools that can help startups to rapidly build, scale, and innovate. This widespread adoption and democratization of machine learning (ML), specifically with the transformer neural network architecture, is an exciting inflection point in technology. With the right tools, startups can build new ideas or pivot their existing product to harness the benefits of generative AI for their customers.

Are you ready to build a generative AI application for your startup? Let’s first review the concepts, core ideas, and common approaches to build generative AI applications.

What are generative AI applications?

Generative AI applications are programs that are based on a type of AI that can create new content and ideas, including conversations, stories, images, videos, code, and music. Like all AI applications, generative AI applications are powered by ML models that are pre-trained on vast amounts of data, and commonly referred to as foundation models (FMs).

An example of a generative AI application is Amazon CodeWhisperer, an AI coding companion that helps developers to build applications faster and more securely by providing whole line and full function code suggestions in your integrated development environment (IDE). CodeWhisperer is trained on billions of lines of code, and can generate code suggestions ranging from snippets to full functions instantly, based on your comments and existing code. Startups can use AWS Activate credits with the CodeWhisperer Professional Tier, or start with the Individual Tier which is free to use.

Figure 1: Amazon CodeWhisperer writes a JavaScript code using comments as the prompt.

The rapidly-developing generative AI landscape

There is rapid growth occurring in generative AI startups, and also within startups building tools to simplify the adoption of generative AI. Tools such as LangChain—an open source framework for developing applications powered by language models—are making generative AI more accessible to a wider range of organizations, which will lead to faster adoption. These tools also include prompt engineering, augmenting services (such as embedding tools or vector databases), model monitoring, model quality measurement, guard rails, data annotation, reinforced learning from human feedback (RLHF), and many more.

Figure 2: Components of the generative AI landscape.

An introduction to foundation models

For a generative AI application or tool, at the core is the foundation model. Foundation models are a class of powerful machine learning models that are differentiated by their ability to be pre-trained on vast amounts of data in order to perform a wide range of downstream tasks. These tasks include text generation, summarization, information extraction, Q&A, and/or chatbots. In contrast, traditional ML models are trained to perform a specific task from a data set.

Figure 3: Demonstrates the difference between a traditional ML model and a foundation model.

So how does a foundation model “generate” the output that generative AI applications are known for? These capabilities result from learning patterns and relationships that allow the FM to predict the next item or items in a sequence, or generate a new one:

In text-generating models, FMs output the next word, next phrase, or the answer to a question.
For image-generation models, FMs output an image based on the text.
When an image is an input, FMs output the next relevant or upscaled image, animation, or 3D images.

In each case, the model starts with a seed vector derived from a “prompt”: Prompts describe the task the model has to perform. The quality and detail (also known as the “context”) of the prompt determine the quality and relevance of the output.

Figure 4: A user inputs a prompt into a foundation model and it generates a response.

The simplest implementation of generative AI applications

The simplest approach for building a generative AI application is to use an instruction-tuned foundation model, and provide a meaningful prompt (“prompt engineering”) using zero-shot learning or few-shot learning. An instruction-tuned model (such as FLAN T5 XXL, Open-Llama, or Falcon 40B Instruct) uses its understanding of related tasks or concepts to generate predictions to prompts. Here are some prompt examples:

Zero-shot learning

Title: \”University has new facility coming up“\\n Given the above title of an imaginary article, imagine the article.\n

RESPONSE: <a 500-word article>

Few-shot learning

This is awesome! // Positive

This is bad! // Negative

That movie was hopeless! // Negative

What a horrible show! //

RESPONSE: Negative

Startups, in particular, can benefit from the rapid deployment, minimal data needs, and cost optimization that result from using an instruction-tuned model.

To learn more about considerations for selecting a foundation model, check out Selecting the right foundation model for your startup.

Customizing foundation models

Not all use cases can be met by using prompt engineering on instruction-tuned models. Reasons for customizing a foundation model for your startup may include:

Adding a specific task (such as code generation) to the foundation model
Generating responses based on your company’s proprietary dataset
Seeking responses generated from higher quality datasets than those that pre-trained the model
Reducing “hallucination,” which is output that is not factually correct or reasonable

There are three common techniques to customize a foundation model.

Instruction-based fine-tuning

This technique involves training the foundation model to complete a specific task, based on a task-specific labeled dataset. A labeled data set consists of pairs of prompts and responses. This customization technique is beneficial to startups who want to customize their FM quickly and with a minimal dataset: It takes a fewer data sets and steps to train. The model weights update based on the task or the layer that you are fine-tuning.

Figure 5: The instruction-based fine-tuning workflow.

Domain adaptation (also known as “further pre-training”)

This technique involves training the foundation model using a large “corpus”—a body of training materials—of domain-specific unlabeled data (known as “self-supervised learning”). This technique benefits use cases that include domain-specific jargon and statistical data that the existing foundation model hasn’t seen before. For example, startups building a generative AI application to work with proprietary data in the financial domain may benefit from further pre-training the FM on custom vocabulary and from “tokenization,” a process of breaking down text into smaller units called tokens.

To achieve higher quality, some startups implement reinforced learning from human feedback (RLHF) techniques in this process. On top of this, instruction-based fine-tuning will be required to fine-tune a specific task. This is an expensive and time-consuming technique compared to the others. The model weights update across all the layers.

Figure 6: The domain adaptation workflow.

Information retrieval (also known as “retrieval-augmented generation” or “RAG”)

This technique augments the foundation model with an information retrieval system that is based on dense vector representation. The closed-domain knowledge or proprietary data goes through a text-embedding process to generate a vector representation of the corpus, and is stored in a vector database. A semantic search result based on the user query becomes the context for the prompt. The foundation model is used to generate a response based on the prompt with context. In this technique, the foundation model’s weight is not updated.

Figure 7: The RAG workflow.

Components of a generative AI application

In the above sections, we learnt various approaches startups can take with foundation models when building generative AI applications. Now, let’s review how these foundation models are part of the typical ingredients or components required to build a generative AI application.

Figure 8: Components of a generative AI application.

At the core is a foundation model (center). In the simplest approach discussed earlier in this blog, this requires a web application or mobile app (top left) that accesses the foundation model through an API (top). This API is either a managed service through a model provider or self-hosted using an open source or proprietary model. In the self-hosting case, you may need a machine learning platform that is supported by accelerated computing instances to host the model.

In the RAG technique, you will need to add a text embedding endpoint and a vector database (left and lower left). Both of these are provided as either an API service or are self-hosted. The text embedding endpoint is backed by a foundation model, and the choice of foundation model depends on the embedding logic and tokenization support. All of these components are connected together using developer tools, which provide the framework for developing generative AI applications.

And, lastly, when you choose the customization techniques of fine-tuning or further pre-training of a foundation model (right), you need components that help with data pre-processing and annotation (top right), and an ML platform (bottom) to run the training job on specific accelerated computing instances. Some model providers support API-based fine-tuning, and in such cases, you need not worry about the ML platform and underlying hardware.

Regardless of the customization approach, you may also want to integrate components that provide monitoring, quality metrics, and security tools (lower right).

Conclusion

In this part of the blog, we learnt various approaches or patterns startups can take to build a generative AI application and the key components involved. In Part 2, we learn how these components map to AWS services, and showcase an example architecture.

Ready to kickstart your generative AI journey? Join AWS Activate, a free program specifically designed for startups and early stage entrepreneurs that offers the resources needed to get started on AWS.

Applications are open for the AWS Global Fintech Accelerator

Sam Edge — Tue, 27 Jun 2023 12:58:36 +0000

Amazon Web Services (AWS) is launching its first Global Fintech Accelerator, giving fintech founders the support and mentorship they need to bring smarter financial services solutions to the market by leveraging the power of AI/ML and the cloud.

Ninety-one percent of financial services institutions are using artificial intelligence (AI) to drive business transformation and change, with 80 percent of fintechs leveraging machine learning (ML) in particular. As the market evolves, AI and ML are standing out as critical enablers of innovation, driving success across the customer journey from user onboarding and workflow automation, to risk management and customer experience.

This change is being driven first and foremost by consumers. Consumer expectations continue to increase, heightened by the hyper-personalized experience they receive from ecommerce companies, social media platforms, and myriad other industries. Financial services are no exception—consumers increasingly expect a faster, more customized, and more convenient way to manage their financial lives. Meeting these evolving expectations is critical to business success, and fintechs need to balance building innovative product experiences with continuing to manage risk and compliance effectively. Artificial intelligence and machine learning are key tools for accelerating success in this new space. We are already seeing successful applications in use cases such as consumer chatbots, digital identification (ID) verification, real-time fraud detection and prevention, and point of sale (POS) credit checks that enable new business models such as Buy Now Pay Later (BNPL).

The opportunity for innovation is vast: open finance, the “API-fication” of financial services, and the explosion of connected devices has made data more accessible than ever. Combined with the democratization of tooling provided by the cloud, AI/ML is no longer limited to a select few; it can be leveraged by anyone and across all stages of maturity. Add in the game-changing impact of generative AI and the potential for reinvention is both broad and deep.

“We are making big bets on the potential of generative AI to significantly change every customer experience out there, especially as it relates to financial services,” said Howard Wright, Vice President (VP) and Global Head of Startups. “In the long term, we expect generative AI to usher in the era of self-driving money, when every consumer with a smartphone can have access to an affordable, trusted financial advisor that will help them improve their financial health. These solutions will leverage what the cloud does best: it democratizes access for everyone.”

With a global infrastructure spanning 99 availability zones and 31 geographic regions, AWS provides the most extensive, secure, and reliable cloud infrastructure in the world. Through our global fintech and financial services teams we help fintech startups access data in a compliant, cost-effective, and scalable way. By launching the AWS Global Fintech Accelerator, we are taking a front seat in the next wave of innovation by accelerating the success of fintech builders from Paris to New York, Buenos Aires to Cape Town, and everywhere in between.

Tailored benefits

The AWS Global Fintech Accelerator will select 150 fintech startups harnessing the power of artificial intelligence and machine learning technology from across North America (NAMER), Europe, Middle East, Africa (EMEA), and Latin America (LATAM). Selected startups will join a six-week remote program between October 2 and November 6, where they will receive 1:1 business and technical mentorship covering funding, scaling, and go-to-market strategies, as well as AWS technical fundamentals, cloud and AI/ML adoption, and much more.

Cohort members will also receive a package of tailored benefits, including up to $25,000 in AWS Activate credits and a range of other benefits provided by accelerator partners across fintech, venture capital (VC), and financial services.

At the end of the six weeks, the top 15 companies in the program will be invited to participate in a Demo Day with a curated list of VCs, financial services institutions, potential customers, and other key market players. They will receive up to $75,000 in additional AWS Activate credits.

Partners

To offer tailored programming for what we expect will be a diverse and wide-ranging group of fintechs, AWS has partnered with key players in the industry, such as NVIDIA, a world leader in AI computing. “Some of the world’s leading fintechs use the full-stack NVIDIA platform to harvest data intelligently and seamlessly through AWS,” said Malcolm deMayo, Global Vice President of Financial Services at NVIDIA. “Through this collaboration, we can help many more startups, especially in our own startup program NVIDIA Inception, to jumpstart their AI and ML journey and bring new financial services to the market.”

Cohort members will also have access to fintech investors’ mentorship, including Bain Capital Ventures. “We’re excited to support AI machine learning within the fintech sector, where we invest,” said Sarah Hinkfuss, Partner at Bain Capital Ventures. “We’ve seen that fintech companies need a long runway from concept, to building, and go-to-market, and are excited to help the companies in this accelerator navigate a tight investment market.”

The accelerator is also supported by Vestbee, one of the largest European platforms for startups, VC funds, accelerators, and corporates.

How to apply

Applications for the AWS Global Fintech Accelerator are open June 27 – August 31, 2023, with limited space available. Startups are encouraged to review the terms and conditions and submit applications here.

The AWS Global Fintech Accelerator gives high-potential seed-stage fintech companies the specialized mentorship, resources, and partnership opportunities they need to better leverage AI, innovate financial services, and meet new customer demands—all while creating powerful solutions in the cloud. Apply online today.

A look back at the first year of the AWS Impact Accelerator

Keely O'Neill — Mon, 26 Jun 2023 12:57:36 +0000

Last year, Amazon Web Services (AWS) launched the AWS Impact Accelerator for startups led by underrepresented founders—giving high-potential, pre-seed startups the tools and knowledge to reach key milestones, such as raising funds or being accepted to a seed-stage accelerator program, while creating powerful solutions in the cloud.

Over the past year, we’ve received thousands of applications from startups building for industries like artificial intelligence/machine learning (AI/ML), fintech, healthcare and life sciences, many of whom came to us through our partners such as Goodie Nation, Advancing Women in Tech, Visible Hands, and Startupeable.

So far, AWS has provided $1.25 million in AWS credits through AWS Activate, enabling startups to build and test their products alongside AWS experts and AI/ML technical teams helping them to adopt features like Amazon SageMaker, Amazon CloudWatch, Amazon Lightsail, and more. Startups also got the chance to accelerate their business alongside AWS Partners, including speaker coaches, to pitch their businesses to 50+ investors at Investor Day, the final event culminating the end of each cohort.

Members developed a strong community where they continue to support each other through funding rounds, growing their teams, and launching new features. This is just the beginning, but we’re happy to celebrate some of their collective wins today.

Black Founders cohort group photo

Since the first cohort—Black Founders—ended in July 2022, members have launched their products, raised millions in funding, hired new team members, increased their revenue by up to 10x, and participated in accelerators run by Techstars, a16z, Expedia, Comcast NBC Universal, Seed to Harvest, and more. Their pitch practice paid off—Alyse Dunn, Founder and CEO of CareCoPilot, won the Fearless Fund pitch competition and Toby Egbuna, Founder and CEO of Chezie, won the Startup Battle Atlanta pitch competition.

To see a behind-the-scenes look at the experiences of members of the Black Founders cohort, check out Making an Impact—a short film documenting the journey of ChurchSpace and Oben Health.

Members of the second cohort, Women Founders, have already made a big impact in the seven months since their cohort ended in December 2022. Following a successful Investor Day at the AWS Startup Loft in San Francisco, members have since been accepted into Techstars accelerator programs—Standd, an AI-enabled knowledge discovery platform for legal teams, even earned Geekwire’s “Best Pitch” of the night for Techstars Seattle.

Women Founders cohort week 8 group photo

Startups in the cohort are still early in their fundraising journeys, but Native-founded and -led fintech startup, Totem, raised a $2.2 million pre-seed round, according to Tribal Business News. With this round, Amber Buker, co-founder and CEO of Totem, joins an exclusive group of only a handful of indigenous women founders to successfully close multi-million-dollar capital raises. We can’t wait to see what Totem and the rest of the companies accomplish in the years to come.

Inclusive Journeys Founder, Parker McMullen Bushman (L) with Amber Buker (R)

The AWS Impact Accelerator Latino Founders cohort was just announced in May 2023, and 20 Latino-led startups made their way to Amazon HQ in Seattle for week one of programming. On June 26, they joined founders from the previous cohorts and members of the AWS Startups team on stage for the Nasdaq Closing Bell Ceremony.

Latino Founders cohort group photo

What’s next for the AWS Impact Accelerator

Despite years of industry effort to change the dynamics of inclusion and diversity in tech, the majority of funding still skews significantly toward startups founded by straight, white men. Venture capital (VC) firm Backstage Capital found that still only 1% of venture-backed founders are Black, 1.8% Latino, and 9% women.

“I am restless,” says creator of AWS Impact Accelerator and Head of Startup Marketing Programs at AWS, Denise Quashie. “There’s still so much work to be done. I know how many deserving underrepresented founders need access to these resources to level the playing field.” Her message to these founders is, “We see the work you’re doing, and we will remain restless until we can help you unlock resources and scale your startup, too.”

Future cohort dates have not yet been announced, but stay tuned to this blog and our social channels (LinkedIn, Twitter, Facebook, and Instagram) for news and announcements about the AWS Impact Accelerator and your next chance to apply.

Want to learn more about our previous cohorts?

Selecting the right foundation model for your startup

Aaron Melgar — Thu, 22 Jun 2023 12:50:51 +0000

When startups build generative artificial intelligence (AI) into their products, selecting a foundation model (FM) is one of the first and most critical steps. A foundation model is a large machine learning (ML) model pre-trained on a vast quantity of data at scale resulting in a model that can be adapted to a wide range of downstream tasks. Model selection has strategic implications for how a startup gets built: Everything from user experience and go-to-market, to hiring and profitability, can be affected by selecting the right model for your use case. Models vary across a number of factors, including:

Level of customization – The ability to change a model’s output with new data ranging from prompt-based approaches to full model re-training
Model size – How much information the model has learned as defined by parameter count
Inference options – From self-managed deployment to API calls
Licensing agreements – Some agreements can restrict or prohibit commercial use
Context windows – How much information can fit in a single prompt
Latency – How long it takes for a model to generate an output

Following are some of the most impactful aspects to consider when selecting a foundation model to meet your startup’s needs.

Application-specific benchmarks

As startups evaluate the performance of different models for their use case, a critical step in the process is establishing a benchmark strategy, which helps a startup quantify how well the content that a model generates matches expectations. “There are a large number of models out there, ranging from closed source players…to open-source models like Dolly, Alpaca, and Vicuna. Each of these models have their own tradeoffs — it’s critical that you choose the best model for the job,” explains Noa Flaherty, chief technology officer (CTO) and co-founder of Vellum. “We’ve helped businesses implement a wide variety of AI use cases and have seen first-hand that each use case has different requirements for cost, quality, latency, context window, and privacy.”

Generalized benchmarks (such as Stanford’s Holistic Evaluation of Language Models) are a great starting point for some startups because they help prioritize which foundation models to start experimenting with. However, generalized benchmarks may be insufficient for startups that are focused on building for a specific customer base. For example, if your model needs to summarize medical appointments or customer feedback, the model should be evaluated against how well it can perform these specific tasks.

“To do custom benchmarking, you need a workflow for rapid experimentation – typically via trial and error across a wide variety of scenarios. It’s common to over-fit your model/prompt for a specific test case and think you have the right model, only for it to fall flat once in production,” Noa advises. Custom benchmarking may include techniques such as calculating BLEU and ROUGE scores; these are two metrics that help startups quantify the number of corrections that are necessary to AI-generated text before giving it final approval for human-in-the-loop applications.

Quality metrics and model evaluation are critical, which is why Noa founded Vellum in the first place. This Y-Combinator backed startup focuses their product offerings on experimentation: Per Noa, “The more you can compare/contrast models across a variety of cases that resemble what you’ll see in production, the better off you’ll be once in production.”

Smaller, purpose-built models are on the rise

Once quality benchmarks have been established, startups can begin to experiment with using smaller models meant for specific tasks, like following instructions or summarization. These purpose-built models can significantly reduce a model’s parameter count while maintaining its ability to perform domain-specific tasks. For example, startup GoCharlie is partnered with SRI to develop a marketing-specific multi-modal model with 1B parameters.

“One-size-fits-all models will never truly solve an end user’s needs, whereas models designed to serve those needs specifically will be the most effective,” explains Kostas Hatalis, the chief executive officer (CEO) and co-founder of GoCharlie. “We believe purpose-built models tailored to specific verticals, such as marketing, are crucial to understanding the genuine requirements of end users.”

The open-source research community is driving a lot of innovation around smaller, purpose-built models such as Stanford’s Alpaca or Technology Innovation Institute’s Falcon 40B. Hugging Face’s Open LLM Leaderboard helps rank these open-source models across a range of general benchmarks. These smaller models deliver comparable benchmark metrics on instruction-following tasks, with a fraction of the parameter count and training resources.

As startups customize their models for domain-specific tasks, open-source foundation models empower them to further customize and fine-tune their systems with their own datasets. For example, Parameter-Efficient Fine-tuning (PERT) solutions from Hugging Face have shown how adjusting a small number of model parameters, while freezing most other parameters of the pre-trained LLMs, can greatly decrease the computational and storage costs. Such domain adaptation based fine-tuning techniques are generally not possible with API-based proprietary foundation models which can limit the depth to which a startup can build a differentiated product.

Focusing usage on specific tasks also makes the foundation model’s pre-trained knowledge across domains like mathematics, history, or medicine, generally useless to the startup. Some startups choose to intentionally limit the scope of foundation models to a specific domain by implementing boundaries, such as Nvidia’s open-source NeMo Guardrails, within their models. These boundaries help to prevent models from hallucination: irrelevant, incorrect, or unexpected output.

Inference flexibility matters

Another key consideration in model selection is how the model can be served. Open-source models, as well as self-managed proprietary models, grant the flexibility to customize how and where the models are hosted. Directly controlling a model’s infrastructure can help startups ensure reliability of their applications with best practices like autoscaling and redundancy. Managing the hosting infrastructure also helps to ensure that all data generated and consumed by a model is contained to dedicated cloud environments which can adhere to security requirements set by the startup.

The smaller, purpose-built models we mentioned earlier also require less compute intensive hardware, helping startups to optimize unit economics and price performance. In a recent experiment, AWS measured up to 50% savings in inference cost when using ARM-based AWS Graviton3 instances for open-source models relative to similar Amazon Elastic Compute Cloud (EC2) instances.

These AWS Graviton3 processors also use up to 60% less energy for the same performance than comparable Amazon EC2 instances, which helps startups who are considering the environmental impacts of choosing power hungry inference hardware. A study from World Economic Forum detailed the energy consumption of data centers. Once considered an externality, environmental implications have risen to top of minds of many and AWS enables startups to quantify their environmental impact through offerings such as Carbon Footprint Reporting, which helps companies compare the energy efficiency of different hardware selections.

Conclusion

Wherever your startup is in its generative AI journey—getting the infrastructure ready, selecting a model, or building and fine-tuning–AWS provides maximum flexibility for customers. Amazon Bedrock, a fully managed service, gives you access to foundation models from leading foundation models including Amazon’s own Titan family of models, available via a fully managed API. Amazon SageMaker JumpStart is self-service machine learning hub. It offers built-in algorithms, pre-trained foundation models, and easy-to-use solutions to solve common use cases for customers like fine-tuning their models or customizing their infrastructure.

Check out these generative AI resources for startups building and scaling on AWS ?. Need help deciding which model or solution to choose? Want to work with AWS to offer your own model or algorithm? Reach out to our team today!

How startup CFOs can integrate the cloud into their long-term success strategy

Kate Radinovic — Wed, 21 Jun 2023 15:42:51 +0000

You are reading the fourth installment of our thought leadership spotlight, “The evolving role of the startup CFO.” This series features perspectives from prominent players in the startup ecosystem.

Imagine you’re a CFO leading the quarterly forecasting meeting for your startup. Your CEO points out a line item in your spending plan–cloud computing services–and asks, “How are you predicting this spend?”

As executives assess the impact of cloud adoption on bottom-line costs and top-line revenue growth, it’s more important than ever for CFOs to understand the cloud and develop metrics to effectively allocate resources and objectively measure return on investment. Equally important, CFOs need to effectively forecast cloud spend and communicate the business’s plans to stakeholders at every level of the company. To explore this, we sat down with Danel Dayan, investor at Battery Ventures, a global, technology-focused investment firm to discuss how CFOs can strategize and plan for the cloud through each stage of company maturity.

“You’ve got to figure out the core metrics you’re optimizing for based on your current stage of development,” says Danel. In his role, Danel focuses on early stage and growth-equity investments in cloud infrastructure, big data, security, and next-generation enterprise applications. “As the CFO of a fast-growing company, you have to have a plan to eventually grow into gross margins of 75-80% annual recurring revenue (ARR). You should be adapting your roadmap with the destination in mind.”

Startup phase

The startup phase includes new businesses earning between $0-5 million ARR. During this phase, for tech startups, it’s common for a majority of resources to be funneled into research and development (R&D). As many early stage companies are not yet able to bring a CFO into the fold, Danel warns startup leaders not to let R&D become a vacuum.

“It’s actually okay to be somewhat over-provisioned as you’re deploying a product early on,” he says. “But there’s this assumption that as long as you’re investing in R&D and product, it’s okay to spend. Up to this point, CFOs have rationalized cloud infrastructure and other R&D expenses as necessary to optimize overall costs, but we can think more critically.”

Danel recommends that early stage startups focus on accelerating product velocity, making sure that the team has adequate resources and budget to take their product to market. Then, they can develop solid performance indicators to ensure that the investment is paying off, such as measuring logo adoption, customer acquisition, and ultimately ARR growth. “[At Battery], we’re very objective around what makes a successful sales or marketing team. Are they hitting quota? How many leads are they able to close? What are their on-target earnings relative to their quota? How much capacity do they have?”

A modern CFO should feel comfortable navigating conversations with their chief product officer, chief technology officer, and head of R&D and leading board-level conversations about R&D investments, and how they’re driving business outcomes such as: percentage of the product completed, expected ARR to be generated, or ratios on new product development versus product maintenance.

Scale-up phase

The scale-up phase begins when a company enters the $5-15 million ARR territory. In this phase, the product is established and reliably profitable. As the company founders set their sights on full maturation, they typically begin to build a more robust executive team to help steer the ship. That means a new level of scrutiny on company expenses that CFOs need to be prepared to navigate.

“In the last few years, we’ve witnessed the trend of startups hiring as much talent as possible versus the right talent, for the right stage, for the right opportunity,” says Danel. Hiring hastily is not only a potential blow to cash flow, but it can create rifts within the company culture. “Being very intentional about team-building is important. We’re seeing today the effects of companies that rushed hiring to meet the demands of their scaling business. And now they’re having to right-size their team, which impacts morale.”

Cloud-based services can be a boon to a startup in the scale-up phase, taking on some of the heavy lifting that comes with scaling. The flexibility the cloud offers lets CFOs commit company dollars when necessary and scale up and down as needed. A great scale-up CFO focuses on optimizing investments to reach the profitability, retention, expansion, and ARR metrics for the business.

Maturity

A startup reaches maturity once it crosses the $15 million ARR threshold. At this point, the company is likely functioning with a full board to advise on resource allocation. It’s crucial at this stage that the CFO is able to communicate the value of their cloud strategy and predict cloud spend, both at present and down the line.

“Ultimately, what investors want is predictability,” says Danel. He explains that CFOs should consider how to build a foundation of trust and predictability when reporting to internal and external stakeholders. “Where a great CFO shines is being able to dial in a business’s budget or forecast quarter after quarter. It doesn’t have to always be positive data, but it needs to be predictable.”

Among their many functions, CFOs of mature startups act as liaisons, communicating between the functional leaders of a company (chief marketing officer, head of sales, chief product officer) and the high-ranking execs to build a model that supports growth and keeps costs under control. Danel says that tension is necessary to balance. “Part of a CFO’s role is to maintain a healthy bit of friction with both sides. Where the CFO really comes into play is providing a counterpoint. [CFOs] manage expectations so that the bigger picture never gets lost.”

No matter what stage of maturity a company is in, Danel says the important thing for CFOs to keep in mind is the relationship between cloud expenses and the benefits they provide. “Those two things go hand-in-hand. You can’t think of them in isolation.” As cloud-based solutions change the game for startups at all levels of maturity, it becomes more standard for CFOs to understand the cost implications and how they work for or against their overall growth strategy. As Danel puts it, an effective CFO is one that can communicate the value of such services, incorporate them into a cohesive financial roadmap, and manage the healthy bit of tension that comes along with building a business.

This article is just one in a series on the evolving role of the CFO. Find the other installments in the series:

AarogyaAI uses AI/ML on AWS to precisely diagnose antimicrobial resistance

Praapti Jayaswal — Fri, 16 Jun 2023 11:35:24 +0000

Founders are familiar with the startup dream. Built on an accumulation of life experiences, victories, defeats, data, iterations—and sometimes luck—this dream takes a startup on the journey from “I’ve got this great idea” to proving what’s possible.

AarogyaAI, a healthcare and life sciences (HCLS) startup in India, is built on a dream to “use the research that my co-founder and I did in academia and move the needle on human health,” says chief executive officer (CEO) Dr. Praapti Jayaswal. AarogyaAI uses artificial intelligence and machine learning (AI/ML) to rapidly diagnose drug resistance in patients caused by bacterial, fungal, and viral pathogens. This allows clinicians to make data-driven treatment decisions and prescribe drugs that effectively treat and increase health outcomes for patients.

Praapti and her co-founder, Avlokita Tiwari, the chief technology officer (CTO), began their friendship in 2012 as carpool buddies during the 30-mile drive from home to their lab at the Translational Health Science and Technology Institute in Faridabad, India. Praapti was working on her PhD in tuberculosis research and Avlokita was a junior research fellow.

Their friendship continued into 2019, when Avlokita told Praapti, “I’m looking for jobs.” Praapti responded, “Don’t look for jobs—come build this company together.’” Without hesitating, Avlokita simply answered “Yeah.”

Dr. Praapti Jayaswal and Avlokita Tiwari, co-founders of AarogyaAI

With that, AarogyaAI was founded. The company celebrated its 4^th birthday in May of 2023 and today has nine clinical and commercial partnerships. They are running six infectious disease pipelines on the Amazon Web Services (AWS) cloud, as well as their repository of nearly 20,000 pathogen genomes.

Proving what’s possible in modern healthcare

“It frustrated us that people were dying of curable diseases. In 2018, tuberculosis was the most lethal infectious disease, even though there were 19 drugs to cure that bacterial infection,” says Praapti. “We saw AI being used to make people richer, but not to make sick people healthier.”

Praapti and Avlokita decided to change this by using AI to make precision diagnosis for antimicrobial resistance accessible at point of care. AarogyaAI allows clinicians to input the genomic sequencing of the patient’s specimen into AAICare, an application built on AWS. Both Praapti and Avlokita believed that everyone can and should have access to effective treatment. “Our vision is making genomics available at point of care. Computing on the AWS cloud takes us one step closer to driving data-driven treatment decisions for positive outcomes at a population scale,” explains Praapti.

AAICare quantitatively and qualitatively identifies all of the pathogens present, as well as the drug susceptibility of those pathogens. Additionally, AarogyaAI’s machine learning models, which they build on Amazon SageMaker, search for and predict mutations within the specimen that contribute to drug resistance.

“Technology has evolved over the decades, but it’s rarely being used in clinical practice,” says Praapti. “We are changing that.”

Strategic investment from Illumina, the largest sequencing hardware company globally, propels AarogyaAI’s mission.

AarogyaAI is proving that proactive healthcare can become as normal as the reactive healthcare that is commonly practiced today. “We are demonstrating that the current genomics of pathogens and superbugs can predict what their future trends will be,” says Praapti. “When a drug comes out, it’s already 10 years too late. We need to apply technology to be able to create drugs preemptively, so that when a virus such as SARS-CoV-2 arises to wreak havoc across the globe—8 billion people—we’re not at the mercy of that invisible thing.”

To biologically and scientifically understand the evolutionary trends of pathogens, “Intelligent genomics is the crux of our work where we combine AI with genomics,” explains Avlokita. “This helps in making treatment decisions today with the vision of extracting relevant information on pathogen evolution, thereby guiding public health policy-making for pandemic preparedness.”

Building their startup on AWS

Praapti and Avlokita decided that to build a secure, reliable, and scalable healthcare application, they would go all-in on AWS.

Aarogya AI uses AWS managed services like Amazon Sagemaker, Amazon Relational Database Service (Amazon RDS), and Elastic Load Balancing to complete tasks which would otherwise be cumbersome for engineering teams to set up and manage. This allows AarogyaAI to focus more on feature development of their product, instead of focusing on operating on a technology data center and the overhead that comes along with that.

“The best part about using AWS is how seamless it all is,” explains Avlokita. “It has been an amazing learning experience, to figure out all of the different features on AWS and how they can support our goals. We’ve been very happy with AWS.”

To accelerate their startup’s growth, AarogyaAI applied for and was accepted into AWS ML Elevate, an India-based program that helps artificial intelligence and machine learning (AI/ML) startups to showcase their innovations while providing mentorship and access to VC channels.

Training and deploying machine learning models

Machine learning models are key to AarogyaAI’s ability to accurately analyze patient specimens and predict future trends. Amazon SageMaker is an artificial intelligence and machine learning solution that helps their data science team to build, train, and fine-tune the machine learning models while providing complete control and visibility. “We rely on Amazon SageMaker to build our AI/ML models and deploy them,” says Avlokita. AarogyaAI trains their Amazon SageMaker models on global and local genomic datasets available both publicly and generated on site, keeping the idea of translating research into real-world application in mind.

“When our product lead joined the company, he had no experience on AWS. He knew data science and AI/ML algorithms, but he didn’t know how to use AWS,” says Avlokita. “Amazon SageMaker is so user friendly that he was able to jump right in and figure out how to build machine learning solutions.”

Computations for the machine learning models run using Amazon Elastic Compute Cloud (Amazon EC2), which offers secure and reliable compute capacity on demand. For object storage, AarogyaAI uses Amazon Simple Storage Service (Amazon S3).

Optimizing cloud costs to reinvest in their business

As the next step in their product development, AarogyaAI wanted to make their AI algorithms more robust and able to analyze more pathogens. “We used AWS to run machine learning experiments. It was quick and efficient,” says Avlokita. “We were very happy with the results. AWS allowed us to take away what we needed from these experiments and learn how to tailor our algorithms to include more pathogens.”

As AarogyaAI experimented and learned how to expand their product, the amount of computation they used increased and affected their AWS bill. However, as members of AWS Activate—a program that helps startups to build and scale—they were able to apply $100,000 in AWS credits towards their bill.

Additionally, “AWS in both India and the US were so generous in letting us know about different aspects of cost optimization. We were able to cut down our bill by 38% almost immediately,” says Avlokita. “The way that AWS helped us to deal with cloud costs is something we’re very happy with. Now, we can keep building.”

Looking ahead at AarogyaAI and healthcare on the cloud

Both Praapti and Avlokita are certain that the future of more equitable and effective healthcare is connected to technology. “Front and center, healthcare is going to rapidly become more tech-driven,” says Avlokita. “It’s important to us that we make products with the capability and scalability to adapt to new tech as it arrives.”

As an early-stage HCLS startup, AarogyaAI leverages the insights of peers, mentors, and experts to navigate the intricacies of entrepreneurship and biotech innovation.

For others that are building at this rapidly emerging intersection of healthcare and technology, Praapti advises, “The biggest thing I can share with other founders is: Don’t wait for anything. Dive straight into the deep end.” She explains, “It’s never going to be the perfect time to reach out to someone or send that email or make that call or start that business or file that patent—whether you make the right call or the wrong call, you’re going to come out net positive.”

As the next steps in their startup journey, AarogyaAI plans to deploy their product more widely in the market, particularly in the US. “We’ve had super amazing support from AWS in India, Singapore, Asia Pacific, and in the US,” says Praapti. “We are excited to begin to co-commercialize with AWS.”

“In 2012, when Avlokita and I used to carpool to the lab, we’d talk about our wildest dreams, like clinical trials becoming redundant using computational biology,” laughs Praapti. “ With AarogyaAI and AWS, our wildest dreams are slowly becoming a reality.”

Curious about how AWS can help kick start your startup? Join AWS Activate to build and scale your startup with the right resources at the right time.

Check out more AIML and HCLS startups building and scaling on AWS ?:

How C2i Genomics builds on AWS to transform cancer care

Boris Oklander — Wed, 31 May 2023 15:43:29 +0000

Healthcare and life sciences (HCLS) startups recognize that technology is an impactful vehicle for advancing human health at speed and scale. More importantly, HCLS startups are working to do something about it. C2i Genomics, founded in 2019, is one such startup: C2i Genomics is building a whole genome intelligence platform to improve cancer monitoring.

Using artificial intelligence (AI) and machine learning (ML) solutions, C2i Genomics’ platform analyzes sequenced genome data to detect the tumor burden of cancer patients via a simple blood test. Its cancer surveillance system can track tumors on the genomic level, giving extensive insight into a patient’s cancer treatment journey. The platform can eliminate the reliance on imaging technology while improving the accuracy of cancer screenings and treatment recommendations.

The biggest benefit of this approach is that it allows for “high-precision personalized medicine,” says Boris Oklander, co-founder and chief technology officer (CTO) of C2i Genomics. “The treatment the patient gets is not just arbitrary protocol, but tailored for their specific case.”

Overcoming challenges using AWS

Boris Oklander, C2i Genomics co-founder and CTO

As they built their whole genome-based intelligence platform, the team at C2i Genomics quickly discovered a significant technological challenge: large data volume. Each blood sample collected from a patient translates into files that’s roughly 100 gigabytes. Patients may have several samples taken throughout the course of their cancer diagnosis and treatment. As their company scaled, data volumes expanded rapidly, and C2i Genomics—despite being a young company—began working with multiple petabytes of data. Making that much data available for processing and analysis is a formidable task for any company.

Beyond that, C2i Genomics faced a complicated legal landscape. Genomic data is sensitive material, subject to privacy laws worldwide, but the specific regulations governing its use can vary from country to country.

By working with Amazon Web Services (AWS), C2i Genomics found assistance with both of these complicating factors. It’s a collaboration that enabled C2i Genomics to manage potential issues efficiently and cost effectively, and it put the company on track to impact healthcare worldwide.

“Utilization of the AWS platform was really a key factor in our success,” says Boris.

Activating success

To help jumpstart their business, AWS provided C2i Genomics with credits through the AWS Activate program. Boris notes that these credits were an essential resource during the company’s early stages. C2i Genomics put its credits toward data storage and computation solutions—crucial components of its genome intelligence platform that would otherwise prove costly for an early-stage startup.

AWS also introduced C2i Genomics to the BeyondBio SCALE startup accelerator program, a program with AstraZeneca and several other organizations. As a participant in BeyondBio SCALE, C2i Genomics received expert guidance from leaders in the healthcare and life sciences sector. The team gained critical insight into the process of scaling up development and deployment of cloud-based medical services, data privacy and compliance, and how to avoid common mistakes that can trouble emerging companies in the field.

After C2i Genomics’ platform went through the due diligence process, AstraZeneca selected the platform to be used in their own labs, cementing a relationship that AWS helped facilitate.

Optimizing costs and solutions

Creating a pioneering diagnostic solution can be an expensive undertaking, even with the early credits assistance. “The costs associated with processing these volumes of genomic data are huge,” says Boris. “And in the current economic environment, we have become much more sensitive to the costs.”

To better manage their expenses, C2i Genomics worked closely with the solutions architect at AWS to identify and explore three key areas of cloud optimization.

Automatically optimizing storage costs

The first area of optimization involved using Amazon S3 Intelligent-Tiering, which monitors data for changes in access patterns and automatically moves data to the most cost-effective access tier. This can produce significant savings for companies such as C2i Genomics, who have variable data access patterns.

Choosing the AWS solutions to meet their use cases

Next, C2i Genomics and AWS worked together to find the most efficient AWS solutions for their use cases. To allow C2i Genomics’ researchers to launch experiments and evaluate algorithms results quicker, they implemented Amazon Managed Workflows for Apache Airflow (Amazon MWAA). C2i Genomics also chose to use Amazon FSx for Lustre for more efficient storage.

During the course of maximizing efficiencies, the team recognized an opportunity to optimize costs by transferring some on-demand instances of Amazon Elastic Compute Cloud (Amazon EC2) to spot instances—in some cases, crafting solutions tailored to C2i Genomics’ unique data needs.

Securing genomic data on the cloud

The third major optimization challenge involved with uploading personalized genomic data to cloud storage—a centerpiece of C2i Genomics’ platform. C2i Genomics needed to assure their customers that they handle their data responsibly while navigating a complex global regulatory environment. “We needed to work both on the technological side but also on legal to make this happen,” says Boris. “The help of the AWS team was really crucial here.”

By adopting Amazon GuardDuty, AWS Config, AWS Systems Manager, and third-party solutions, C2i Genomics succeeded in building a regulated environment in which to host their genomics platform. The alternative would have required deploying their platform on-premises for each customer rather than via the cloud—a costly, non-scalable option. By collaborating with the AWS team, C2i Genomics was able to turn its vision of a global diagnostic ecosystem into a reality.

Using Amazon Omics

A recent product launch meant that AWS was uniquely positioned to assist C2i Genomics in their endeavor: In 2022, AWS launched Amazon Omics, a service designed specifically for companies in the healthcare and life sciences space. Omics provides a venue for storing, querying, and analyzing biological data, including genomics. By bringing all of this data onto one accessible platform, Omics fosters collaboration among teams and expedites scientific innovation.

By including Omics in their tech stack, C2i Genomics is able to rely on AWS for its genomic data storage and processing infrastructure. Omics is developed to specifically handle the kinds of workloads required to generate insights from huge volumes of genomic data, relieving C2i Genomics of a significant amount of engineering stress. C2i uses Omics as an on-demand service—it’s available immediately as needed, but the startup’s engineers don’t have to worry about maintaining this expansive layer of infrastructure on their own.

In addition, Omics is General Data Protection Regulation (GDPR) compliant and Health Insurance Portability and Accountability Act (HIPAA) eligible, which allows C2i Genomics to focus on tackling problems in cancer diagnosis and treatment, rather than on regulatory frameworks.

Proving what’s possible with AWS

The impact of C2i Genomics’ platform has the potential to ripple across the global healthcare industry, transforming what it means for someone to receive a cancer diagnosis. It can lead to more personalized treatments and better health outcomes. By monitoring a patient’s progress throughout treatment, a healthcare team can make informed decisions about whether to continue an aggressive form of therapy, switch to another, or discontinue treatments that may no longer be necessary.

The C2i Genomics team celebrates Passover with a toast at their office.

“We’ve the help of AWS, we have achieved a point where we not only have a medical device, but it’s clinical grade,” Boris says. “It’s not only for research; it’s actually ready for providing results in short turnaround times for real patients.”

Similarly, C2i Genomics’ platform offers the ability to detect whether a form of treatment has successfully eliminated a cancer. This can help prevent the need for surgeries that remove organs, which are sometimes performed preventatively due to a lack of available information about how far a cancer has spread.

C2i Genomics’ ambition is nothing less than making this kind of personalized medicine available to all patients across the globe. It aims to scale its platform so that any lab with the right equipment can take advantage of these innovations in cancer care, getting diagnostic screening results in record time. The company knows that doing this will require designing its service and implementation in close collaboration with AWS, a prospect Boris looks forward to.

“On the AWS side, they understand that we are on a mission to transform cancer treatment,” says Boris. “It’s a unique position where C2i Genomics can utilize AWS’ advanced cloud-based technology to save lives.”

Ready to begin your startup journey? Join AWS Activate to build and scale your startup with the right resources at the right time.

Learn more about how startups are using AIML solutions on AWS:

Why Guild CFO Chris Garber believes lifelong learning is the key to startup success

Kate Radinovic — Tue, 30 May 2023 17:07:37 +0000

You are reading the third installment of our thought leadership spotlight, “The evolving role of the startup CFO.” This series features perspectives from prominent players in the startup ecosystem.

The ed tech and skilling industry is booming. Businesses and investors are embracing job training and upskilling as strategies to create more adaptable and resilient workforces. But even in a crowded field, Guild stands out: the startup tripled its valuation in 2021, attracting key investors like Bessemer Venture Partners to help the company get to the next level.

Guild offers a sophisticated three-sided marketplace that connects employees who want to advance their skills and achieve career mobility, educational institutions that serve adult learners, and employers who want to invest in their employees’ growth. Guild’s proposition is that every person is a lifelong learner, and that businesses and people will benefit when employers invest in their employees by providing them access to the tools they need to grow.

So, it’s no surprise to hear Chris Garber, Guild’s CFO, sum up the company’s internal ethos: “it’s all about a learner’s mindset.” After sitting behind the CFO desk for a year and a half, Chris is the first to attest to how much he’s learned about steering the finances of a dynamic company like Guild, which must continue to grow while navigating a rapidly evolving landscape.

“The business is changing. The market is changing. The realities for the company are changing. And so, it is really about having your finger on the pulse of all of those dynamics,” Chris says.

Process, not structure

Chris is well aware that the addition of a “numbers guy” to the leadership team can provoke some anxiety about new restrictions or increased paperwork, and that people might yearn for the earlier, less fettered days in the startup’s lifecycle.

“There’s a stage of a company where you don’t need a CFO,” Chris acknowledges, “and another stage where you need a CFO, but you need them to perform a really specific controllership function. And then over time, what you need from that CFO really expands.”

Chris explains that having the right CFO adds a certain level of sophistication and maturity, but should not detract from the company’s ability to innovate, experiment, and change. “Some degree of process actually helps as a catalyst and accelerant to the business,” Chris says. “Good process helps you go faster with more volume. Bad process becomes bureaucratic and slows you down.”

A great startup CFO should not impose a rigid hierarchy. Instead, Chris sees the CFO’s role as helping the existing elements function better together. On many of Guild’s teams, Chris says, information flows laterally rather than too strictly along reporting lines. “We’re optimizing for nimbleness, for context, and for the ability to react as a team, rather than wait for somebody to tell you what to do next.”

A coach, not a scorekeeper

“Now I don’t think about it as, ‘The CFO’s only job is to make sure the books are right,’” Chris says. However, earlier in his career, he says he was much more concerned about holding the business to account against the balance sheet. But a meaningful conversation with the CEO of the company where he worked before Guild changed his mindset. “He said, ‘I need finance to be a coach, not a scorekeeper,’” Chris recalls. “Now, I see the CFO’s role as a guide and a catalyst for helping all the different parts of the business get better.”

From the cross-functional vantage of finance, it’s easier to monitor the different moving parts of the company and see connections that, perhaps, other team-members aren’t yet aware of. “You’re one of the few places where those kinds of dots get connected,” Chris reminds his fellow CFOs. “Our job is to shine light on those things and help each part of the company prioritize and focus on the most important things.”

Speaking many languages

“For a modern CFO to make sense of that high-level view,” Chris says, “it’s essential to be conversant in every aspect of the business’s operations in order to connect strategy and execution. You’re never done learning about the business, about how the organization works.” Chris notes that the best preparation for a startup CFO, along with solid financial bonafides, is deep experience with the non-finance side of the business. “You need to learn the language of engineering. Of sales. Of professional services. Of marketing and product management.”

Other parts of the company might have to learn to speak finance as well. Chris encourages his team members to see financial tools as “a new language we can bring into how we think about what decisions we make.”

Meanwhile, CFOs—especially at startups—have to pick up another language: Chris calls it “persuasive capability.” CFOs have to either make or drive decisions, and they do that, Chris says, by telling compelling stories and making convincing arguments. These are likely to be quantitatively informed, but they have to appeal at a human level.

Nimble at any scale

Despite stereotypes about moving fast and breaking things, Chris says, there’s less difference than one might think between stewarding the finances of a successful startup and those of a more established company. “If you’re in a dynamic business, the behaviors that contribute to success are actually really similar,” Chris says. “The difference is mostly scale.”

Size brings complexity and breadth, as well as more leadership challenges emerging from larger teams. But at the end of the day, and after a career ranging from fast-growing, earlier-stage enterprise companies to a decades-old, publicly traded services company to Guild, Chris knows that the basic skills—what he calls “the DNA” of managing an innovative company—stay the same. Lifelong learning isn’t about becoming someone else entirely; it’s about building on what you already know to be able to thrive in a new way.

Other posts in this series

AWS announces 21 startups selected for the AWS generative AI accelerator

Kathryn Van Nuys — Wed, 24 May 2023 11:55:47 +0000

AWS is excited to announce the cohort of startups accepted into the global AWS Generative AI Accelerator. The program kicks off May 24th at our San Francisco AWS Startup Loft and closes on July 27th.

Over the course of their 10-week program, participants will receive tailored technical advice, dedicated mentorship, an opportunity to pitch their demos to venture capitalists (VCs) in the AWS network, and up to $300,000 in AWS credits. Critically, they will also have the opportunity to foster lifelong connections with their fellow founders and within AWS.

Our finalists come from various industries, backgrounds, and geographic regions, but all they have one thing in common: they are using generative artificial intelligence (AI) technology to drive unprecedented innovation in their space.

They’re exploring practical solutions to problems such as illiteracy and healthcare burnout and designing tools that drastically reduce time spent on costly, tedious tasks.

No matter their vision, all of these startups are proving what’s possible with generative AI and boldly reinventing applications, data touchpoints, and customer experiences, to name a few.

Backing the upcoming leaders of the generative AI landscape

Startups are the lifeblood of innovation, and AWS is eager to support them in developing incredible generative AI solutions. Many of the AWS Startups team are former founders or VCs, and we embrace this chance to give back to these startups in meaningful, actionable ways.

“Generative AI holds tremendous potential to revolutionize how humans interact with technology and with each other, while democratizing access to new and existing technology in a way that is unprecedented.” says Jon Jones, vice president of compute and AI/ML services at AWS. “Customers are already seeing value in streamlining processes, accelerating product development, and using AI as a trusted companion to increase productivity and better serve their clients.” We are excited to partner with these innovators on their journey to solve some of the world’s biggest challenges.”

Drumroll, please

Please join us in extending a warm welcome to the 21 AWS Generative AI Accelerator program finalists.

Education

Ello

Ello leverages large language models (LLM) and AI solutions to perfectly tailor literacy lessons to each young student they reach. Through interactive reading sessions from real books, Ello becomes a motivational learning companion that transforms children into curious, enthusiastic readers.

Marketing, social, and advertising

Crate

On a mission to create an open internet with no boundaries, Crate invites users to curate a personal, shareable artifact made up of their favorite pieces from anywhere on the web. The team puts AI in the hands of users to help them tell better stories with auto generated images, text, and instant summaries.

q lip

qlip is an AI-powered video highlights generator that helps users grow their social media presence by automatically repurposing long-form videos into short highlights primed for today’s audiences.

OpenAds

OpenAds solves advertising challenges for publishers, consumers, and advertisers by identifying and suggesting ads that match a business’ user experience UX, are tailored to customer advertising and privacy preferences, and keep creative control in the hands of advertisers.

Entertainment and gaming

Leonardo Ai

Leonardo Ai is an AI-driven content production suite tailored for creators across diverse sectors, with a core focus on game development artists. Through the platform, developers can utilize generative AI solutions that integrate with their workflows to unlock their creativity and accelerate content production from months to minutes.

Storia

Built by leading AI researchers and engineers, Storia operates as a creative assistant for rapid film previsualization and production. Story producers can experiment with AI-generated videos, visualize what their product would look like shot in different styles, and build collaborative and comprehensive storyboards in minutes.

Krikey

Krikey uses generative AI to make it easier for creators to breathe life into animations, helping them automate character motion with a variety of 3D avatars, augmented reality (AR) gaming toolkits, and 3D animations. Animations can be seamlessly integrated and exported into the creator’s platform of choice, significantly shortening production time and enhancing the creative process.

Poly

Poly is an AI-enabled infinite design asset marketplace (offering seamless physically based rendering [PBR] textures, illustrations, icons, sounds, and many more) that lets anyone use or generate stunning, 8K high definition (HD) professional design assets in seconds with AI.

Flawless

To counteract the rising on-set production costs and time constraints, Flawless gives artists a suite of cinematic-quality AI-powered tools that allow them to rapidly and affordably iterate, experiment, and refine their content.

Healthcare and life sciences

Knowtex

Knowtex empowers clinicians with voice AI automated note-taking and coding from natural conversation to combat burnout and allow focus on patient care.

Vevo

Vevo is building the world’s first atlas of how drugs interact with patient cells in living organisms at single cell resolution. Vevo’s foundation models trained on this atlas faithfully capture disease biology, enabling generative design of drugs that are more likely to treat disease in humans.

Ordaōs

Ordaōs is a human-enabled, machine-driven drug design company. Their miniPRO proteins help drug hunters deliver treatments that are safer and more effective than traditional discovery methods.

Nosis Bio

Nosis Bio is enabling the future of targeted drug delivery by integrating deep expertise in generative AI and high-throughput biochemistry.

Finance

Theia Insights

Theia Insights leverages the power of AI to synthesize and distill financial data, generating real-time insights beyond human research capability, to inform the investment management community, helping individual and institutional investors make better decisions.

Data and knowledge management

Unwrap

Powered by AI and ML, Unwrap analyzes data from multiple customer feedback channels at scale, providing them with auto-labeling, semantic search, and automatic alerts that strengthen the feedback loop between companies and their customers.

Stack AI

Stack is a no-code interface that helps businesses of all sizes build and deploy AI applications including chatbots, document processing, content creation, and automated customer support in minutes.

Nixtla

Nixtla is building a state-of-the-art disruptive open-source ecosystem that uses AI to unlock scalable, lightning-fast, and user-friendly time series forecasting and anomaly detection.

Wand

Wand enables businesses to sync data from multiple sources to rapidly build collaborative, measurable, and scalable AI solutions. From predictive models to customized LLMs, teams have the power to solve business problems and create value faster than ever before.

Griptape

Griptape’s open source framework and managed service enables developers to enhance LLMs with chain of thought capabilities, creating context-aware conversational, copilot, and autonomous agents.

AI ethics, safety, and security

Bunked

Bunked distinguishes AI-generated content from real content using blockchain technology.

Protopia AI

Protopia AI provides data protection and privacy-preserving AI/ML technologies that specialize in enabling AI algorithms and software platforms to operate without the need to access plain-text information. The company works with enterprises and generative AI/LLM providers to enable maintaining ownership and confidentiality of enterprise data while using AI/ML solutions.

AWS is excited to act as a catalyst for these forward-thinking startups. We continue to build upon the legacy of our previous accelerator programs—such as the AWS Impact Accelerator—to provide founders with the resources, guidance, and networking opportunities they need to scale and succeed.

In the same way AWS democratized the cloud by expanding access to industry-leading technology, we look forward to offering our scale, expertise, and relationships to the next generation of companies at the forefront of generative AI innovation.

Autonomous driving startup TIER IV uses AWS to change the future of mobility

Shinpei Kato — Thu, 18 May 2023 10:31:58 +0000

To successfully disrupt an industry, startups must look to push boundaries, to tackle problems using the latest technology, and to prove what’s possible.

Shinpei Kato is the founder, chief executive officer (CEO), and chief technology officer (CTO).

In the automotive industry, TIER IV is an innovative and disruptive startup that is transforming the vehicle production process and the future of mobility. Founded in 2015 by Shinpei Kato in Japan, TIER IV builds platforms based on open source software—platforms they manage using AWS—that their partners use for building autonomous vehicles.

“We want to reimagine intelligent vehicles,” says Shinpei. “I respect the traditional structure of how vehicles have been mass produced, but it has not changed at all for a hundred years.”

The open-source software that TIER IV uses comes from the Autoware Foundation (AWF), a nonprofit that Shinpei co-founded in 2018. Using this software has been integral to the company’s success. It allows smaller startups like TIER IV to compete with bigger companies that have seemingly infinite resources. Open-source software benefits TIER IV’s customers as well, reducing the cost and delivery time of the product without compromising quality. “We manage the operations of our open-source product using AWS,” says Shinpei.

Shinpei maintains that open-source software creates possibilities for innovation in the field of autonomous vehicles because a wider range of people can experiment with it. “The sky is the limit and anyone can participate,” says Shinpei. “That’s the spirit of open source.”

Building on AWS

Given the benefits of the AWS Activate program for startups, Shinpei says the choice to use AWS for TIER IV’s cloud infrastructure was an easy one. TIER IV has also benefited from AWS’s robust technological offerings, as well as its extensive support network of experts and partners. “There are other kinds of private cloud infrastructures from other companies,” he says. “But if you look at the capabilities of engineers, many of them are more familiar with AWS’s Application Programming Interface (API).” It was easy for the TIER IV team to find engineers who were familiar with AWS’s technology and able to leverage it to its full potential.

The wide range of AWS solutions—from the Internet of Things (IoT) and storage to more compute intensive infrastructure—has helped TIER IV build and scale. The company uses AWS IoT solutions for data collection of vehicles in motion so that they can build remote monitoring and fleet management systems. The streaming data is ingested into durable Amazon Simple Storage Service (Amazon S3) to be consumed by multiple services.

Scalable machine learning instances are key

AWS also helped the TIER IV team to optimize their cloud architecture: “Without support from AWS Startups team for architecture design, we couldn’t have launched our product,” says Shinpei, specifically citing AWS’s help with fleet management, remote monitoring systems, and autonomous driving simulations. Creating simulations with AWS compute services such as Amazon Elastic Kubernetes Service (EKS) allows them to test vital components of autonomous vehicles without operating real vehicles on public roads.

“Simulations can reduce the risk of accidents and reduce the cost of engineering,” says Shinpei. “AWS is helping us develop a high-quality, low-cost, and high-class delivery product.”

An autonomous vehicle simulation created on TIER IV’s platform.

When it comes to the role of machine learning, Shinpei believes that AI is necessary for the future of autonomous vehicles—and working with AWS to leverage Amazon SageMaker is key. “AWS is a one-stop platform for us to build intelligent autonomous vehicles,” he explains.

“AWS scalability is important for us. We can run thousands of instances of machine learning processes and simulation processes in parallel,” says Shinpei. “This capability is key for us to build or mass produce autonomous vehicles in the future.”

Driving towards a successful future for autonomous vehicles

The AWS Partner Programs has also been a game changer for TIER IV. One of the perks of participating in AWS Activate was undergoing an AWS Foundational Technical Review (FTR), a process that validates that AWS Partner solutions are following AWS best practices and identifies technical risks in their solutions. “We went through the technology review and, thanks to significant support from AWS members, finally succeeded in being certified by AWS,” says Shinpei. Undergoing this process resulted in a “Reviewed by AWS” solutions badge, which will help TIER IV garner funding and makes them validated in the AWS Partner Network and enables co-selling and co-marketing initiatives with AWS sales teams.

Looking forward, Shinpei says that AWS’s infrastructure and expertise will become more integral to TIER IV’s ability to scale and innovate.

“We will continue to build the future of mobility,” he says.

Ready to begin your startup journey? Join AWS Activate to build and scale your startup with the right resources at the right time.

Learn more about how startups are using AIML solutions on AWS:

How startups lower AI/ML costs and innovate with AWS Inferentia

Shruti Koparkar — Tue, 16 May 2023 13:47:33 +0000

As a machine learning (ML) startup, you’re probably aware of the challenges that come with training and deploying ML models in your applications (“ML productization”). ML productization is challenging because startups are simultaneously working to achieve high application performance, create a delightful user experience, and manage costs efficiently–all while building a competitive and sustainable startup.

When choosing the infrastructure for their ML workloads, startups should consider how to best approach training and inference. Training is process by which a model is built and tuned for a specific task by learning from existing data. Inference is the process of using that model to make predictions based on new input data. Over the last five years, AWS has been investing in our own purpose-built accelerators to push the envelope on performance and compute cost for ML workloads. AWS Trainium and AWS Inferentia accelerators enable the lowest cost for training models and running inference in the cloud.

AWS Inferentia-based Amazon EC2 Inf1 instances are ideal for startups that wish to run ML inference applications such as:

Search
Recommendation engines
Computer vision
Speech recognition
Natural language processing (NLP)
Personalization
Fraud detection

For training and deploying more complex models such as generative AI models (large language models and diffusion models), your startup may want to check out the new AWS Trainium-based Amazon EC2 Trn1 instances and AWS Inferentia2-based Amazon EC2 Inf2 instances.

In this post, we will cover use cases from two startups–Actuate and Finch Computing–and the success they’ve seen with Inferentia-powered Inf1 instances.

Actuate | Threat detection using real-time AI video analytics | 91% savings on inference costs

Use case: Actuate provides a software-as-a-service (SaaS) platform meant to convert any camera to a real-time threat-detecting smart camera to instantly and accurately detect guns, intruders, crowds, and loitering. Actuate’s software platform integrates into existing video camera systems to create advanced security systems. With Actuate’s artificial intelligence (AI) threat detection software, customers receive real-time alerts within seconds, and can act rapidly to secure their premises.

Opportunity: Actuate needed to ensure high detection accuracy. This meant constantly retraining their models using more data, which took up valuable developer time. Additionally, because they needed fast response times, they depended on GPU-based infrastructure which was cost-prohibitive at scale. As a startup with limited resources, minimizing inference costs and developer time could help Actuate use those resources to build better capabilities and provide more value to its end-users.

Solution and impact: First, Actuate implemented Amazon SageMaker to train and deploy their models. This reduced their deployment time–as measured from labeled data to deployed model–from 4 weeks to 4 minutes. In the next phase, they migrated the ML models across the entire suite of their products from GPU-based instances to AWS Inferentia-based Inf1 instances. This migration required minimal developer involvement as they didn’t need to re-write application code and only needed a few lines of code changes. Actuate saw out-of-the-box cost savings of up to 70% with AWS Inferentia. On further optimization, they reduced inference costs by 91%. This allowed them to use their resources to focus on user experience improvements and fundamental AI research.

Resources: To learn more about Actuate’s use case, you can watch their presentation at reInvent. To get started with a computer vision model on Inf1 instances, visit the Neuron documentation page and explore this notebook for Yolov5 model on GitHub.

Finch Computing | Real-time insights using NLP on informational assets | 80% savings on inference costs

Use case: Finch—a combination of the words “find” and “search”—Computing serves media companies and data aggregators, US intelligence and government organizations, and financial services companies. Its products use natural language processing (NLP) algorithms to provide actionable insights into huge volumes of text data across a variety of informational assets. An example of this is sentiment assignment, which involves identifying a piece of content as positive, negative or neutral and returning a numeric score indicative of the sentiment level and type.

Opportunity: After adding support to their product for the Dutch language, Finch Computing wanted to scale further to support French, German, Spanish, and other languages. This would help existing clients with content in these languages, and also attract new customers across Europe. Finch Computing had built and deployed its own deep learning translation models on GPUs, which were cost-prohibitive to support additional languages. The company was looking for an alternate solution that could allow them to build and run new language models quickly and cost-effectively.

Solution and Impact: In just a few months, Finch Computing migrated their compute-heavy translation models from GPU-based instances to Amazon EC2 Inf1 instances powered by AWS Inferentia. Inf1 instances enabled the same throughput as GPUs, but helped Finch save more than 80% on its costs. Finch Computing supported the three additional languages and attracted new customers. Today all their translation models run on Inf1 and they plan to explore Inf2 instances for new generative AI use cases such as text summarization and headline generation.

Resources: To learn more about Finch Computing’s use case, you can read this case study. To get started with a translation model, visit the Neuron documentation page and see this notebook for MarianMT model on GitHub.

AWS Inferentia for cost-effective, high performance ML inference

In this blog, we looked at two startups who cost-effectively deployed ML models in production on AWS Inferentia, while achieving high throughput and low latency.

Are you ready to get started with Inf1 instances? You can use AWS Neuron SDK, which integrates natively with popular ML frameworks such as PyTorch and TensorFlow. To learn how, please visit the Neuron documentation page and explore this sample model repository on GitHub.

Curious about how AWS can help kick start your startup? Join AWS Activate to build and scale your startup with the right resources at the right time.

Check out how more AIML startups are building and scaling on AWS ?:

Evolutionary architectures series, part 4

Aayzed Tanweer — Thu, 11 May 2023 13:59:22 +0000

The hyper-scale growth stage arrives and the Example Startup team is thrilled.

“Would you like coffee with that?”

“Evolutionary Architectures” is a four-part blog series that illustrates how solution designs and decisions evolve as companies go through the different stages of the startups lifecycle. In this series, we follow the aptly named Example Startup whose idea is to create a “fantasy stock market” application, similar to fantasy sports leagues. They envision holding four “tournaments” over the course of a year.

The third blog post described how the startup began evolving their architecture into one that included tooling such as CI/CD pipelines and infrastructure-as-code, as well as implementing best practices, especially around security and authorization. In part 4, we’ll see Example Startup formalizing their security and backup posture to meet various compliance standards. They also set a data strategy for the organization and explore additional lines of business to diversify their product portfolio.

Using Series B funding to hire, expand, and scale

Things are going as well as they can for Example Startup. They recently closed a Series B round of funding that they expect to fuel some much-needed hiring, expansion, and scaling. With the funding and customer adoption has also come increased competition. Well-established players in the space are beginning to see them as serious competitors, and ramping up their marketing efforts.

Example Startup begins hiring to grow out functional areas and create dedicated teams for site reliability engineering (SRE), platform, analytics, and data science. With the competitive labor market and the startup’s lack of a dedicated human resources department, they once again reach out to their AWS account team to ask about finding technical talent that is proficient with AWS. It turns out that the account team has already helped other startup customers in similar situations by suggesting AWS Partners that can assist with their hiring needs. Soon the startup has emails from multiple vetted candidates with AWS experience in their inbox. Fortunately, many of the interviews turn into job offers and the Startup is able to check hiring off of their to-do list.

The hiring spree leads to a technical problem: Teams are complaining that Example Startup’s platform team is taking too long to set up new accounts for testing purposes and this is stifling innovation. The platform team explains that they don’t have the bandwidth to build individual accounts any time someone gets a new idea for a feature. At this point in their AWS journey, the platform team has a biweekly meeting with their AWS account team and they bring up their dilemma. The AWS solutions architect (SA) recommends setting up a Sandbox dedicated Organizational Unit (OU) within their AWS organization to rapidly provision temporary resources and environments for other teams that want to test new AWS services and features. Additionally, to keep costs under control, the SA recommends using automation to automatically stop resources such as Amazon EC2 instances outside of normal business hours. The platform team at Example Startup follows the SA’s advice. They’re able to quickly spin up accounts for the different teams across the startup and do so in a cost-controlled manner.

Following this, the newly hired SRE team realizes that the startup can be better positioned from an availability and disaster recovery (DR) standpoint. They recognize that the startup is growing rapidly and as it begins to target larger customers, the startup will be faced with more stringent security requirements such as audits and compliance reviews. This meant a need for some infrastructural change. Luckily, a lot of the disaster recovery heavy-lifting was accomplished when Example Startup templatized their infrastructure into Terraform and transitioned into a multi-account architecture in part 3 of the series.

As a first step, the Example Startup team familiarizes themselves with the AWS Resilience Hub to learn more about building resilient applications on AWS. The team still has some outstanding questions so they connect with their AWS account team, who introduces them to a SA with expertise in resiliency. The SA works with Example Startup to specify their recovery time objective (RTO) and recovery point objective (RPO) requirements and do a cost/benefit analysis of different disaster recovery scenarios. After a few calls with the SA and a lot of internal deliberation, the SRE team decides that their RTO/RPO requirements do not call for a multi-region setup as of yet. They make the decision to move their transactional data from Amazon RDS for PostgreSQL into Amazon Aurora for PostgreSQL, primarily for the higher availability it offers as well as the Amazon Aurora Global Database functionality that is important for their production database. The startup also uses AWS Audit Manager to evaluate their adherence to the relevant compliance standards for the upcoming audit. Its automated evidence collection functionality saves them a lot of manual effort.

Building a better customer experience

After sifting through product feedback from various customers, the chief technology officer (CTO) realizes that the dashboarding capabilities offered to traders in Example Startup’s trading application are lacking. Competitors allow users to easily see a visual representation of trading history and performance (compared to other traders) at a per trade level. The CTO flags this as a critical feature gap and the work was assigned to the new analytics team. Additionally, the CTO wants the team to allow traders to generate trading reports at will.

The analytics team has prior experience with Amazon Quicksight so the dashboard component will be straightforward to enhance, including an anomaly detection feature to help traders find specific trades that are outliers. The reporting request is more complex as they do not want to upset the developer teams by running those reports on the production database (nor would that be considered good practice). After consulting the Modern Data Architecture on AWS whitepaper, the analytics team realizes that the best way to go about this is to load the Amazon RDS for PostgreSQL data into Amazon Redshift, a data warehouse. By using Amazon Redshift’s massive parallelism, they’ll be able to run complex aggregations against this transactional data with far less latency and with the added advantage of not bottlenecking the production database. The analytics team is pleased to discover that since Amazon Redshift is built on top of the PostgreSQL engine, they can re-use most of their queries.

Adding a new line of business to the startup

As all of these changes take place, the chief executive officer (CEO) attends a meeting with one of her ex-colleagues at a major trading firm. She learns that there is a dearth of effective traders in the market and this is negatively affecting the trading firm’s hiring pipeline and future projects. The CEO calls up a few of her friends at trading firms who confirm this to be an industry-wide shortage. The CEO starts to form an idea: Example Startup has plenty of traders who perform well. What if Example Startup provides their traders with some sort of cohort-based training which would then feed into a talent pipeline for these trading firms? It would give the traders a chance to enter the job market and the trading firms would have new talent.

Since a key portion of the training would be recommending the correct trades to new trainees, the CEO sets up a call with the data science team to learn how quickly they can build a machine learning (ML) model. Fortunately, some of the data science team has prior experience building, training, and deploying models with Amazon Sagemaker. Since Amazon Redshift is one of the available data sources for SageMaker, the data science team won’t have to setup a complex extract, transform, and load (ETL) pipeline. Members of the data science team who are were less experienced with SageMaker were invited to a SageMaker-focused immersion day by their AWS account team to quickly upskill. Soon they too were on their way to creating training jobs, building accurate models, and deploying the models to endpoints. Anticipating the call from Example Startup’s finance team regarding the increasing compute costs, the data science team did some research and found that they could actually run their training jobs (which account for approximately 80% of their total costs) on spot instances. By doing so, they were able to significantly reduce their costs.

Example Startup’s evolved architecture.

As news of this talent development initiative as a new line of business and revenue stream spreads among the industry, more venture capital (VC) firms began expressing interest in Example Startup–even before the startup expresses intent to start another fundraising round!

A few quarters later, guided by marketing efforts, co-marketing efforts with AWS, and some word of mouth from existing customers, the talent development initiative of Example Startup had grown rapidly. As exciting as the growth in customer base is, more exciting is the fact that the startup is, for the first time since inception, in the green! Recruitment turns out to be a profitable business for the startup and it grows by the day. The executive team realizes this steady profitability stream could fuel the rest of their business. They work diligently on further expansion and innovation plans, excited by the incredible future ahead.

Summary

Over the course of this four-part series, Example Startup cycles through the main stages of a startup from nascency to a matured company. Most startups begin with little more than a bold idea and a dedicated founding team. They build a minimum viable product (MVP) with serverless infrastructure that allows them to test their idea in a way that the founding team can manage on their own.

As the startup acquires paying customers and secures post-seed funding, they move onto the “being onto something” phase, where their architecture evolves and they start to think seriously about scaling, security, and development agility. This means improvements such as using build tools and setting up monitoring and purpose-built databases.

One of the most important lessons during these stages is to engage with the AWS team early and often, even before starting projects, to help evaluate options and save time. Access to both business and technical resources can accelerate timelines and help early stage startup teams achieve their goals.

After the product market fit stage, startups may begin hiring more and focus primarily on scaling – the “to the moon” phase. At this point, they start looking at a multi-account strategy, using service-oriented architecture, caching, and other architectural tweaks to optimize their application and improve the customer experience from a technical perspective.

Finally, they enter the hyper-scale stage where they may begin extensive hiring and build out additional lines of business. From a technical perspective, the startup has invested a lot of time and engineering effort in improving things like security and controls and permissions for their environment. They likely have a codified version of their environment which they can leverage for rapid international expansion and disaster recovery, among other use cases. They’ve also built out a data strategy and are in the process of using analytics and machine learning to better understand their customer and business. This final phase can vary. Some startups will be on the path to an acquisition, whereas others may pursue profitability and market dominance.

Ready to begin your startup journey? Join AWS Activate to build and scale your startup with the right resources at the right time.

Check out all of the Evolutionary Architectures series:

Announcing the startups selected for the AWS Impact Accelerator Latino Founders cohort

Denise Quashie — Tue, 09 May 2023 11:59:47 +0000

As part of our $30 million commitment to provide underrepresented founders with the resources, capital, and community they need to level the startup playing field, we are announcing today the 20 companies that will participate in the AWS Impact Accelerator Latino Founders Cohort. The program kicks off this week in our HQ in Seattle and closes in 8 weeks, with an investor pitching day in New York—including the Nasdaq Closing Bell Ceremony, which is set to foreshadow a successful journey to the Latino-led companies that joined this cohort.

Increasing Latino-led startup funding

Despite representing one in five people in the U.S., Latinos are still significantly underrepresented in venture funding. “This cohort of the AWS Impact Accelerator aims to highlight the viability and ingenuity of Latino-led startups, so the VC community can increase support to these founders,” said Howard Wright, Vice President and Global Head of Startups at AWS. “We’re looking forward to playing an active role in helping these companies turbocharge their growth through access to capital, experts, and all of the innovations that the AWS tech stack has to offer.”

Consistent with previous cohorts, the 20 members of this cohort were selected from a competitive field of applicants which included over 1,100 submissions, and chosen by a diverse committee of AWS startup experts based on the strength of their idea, technical readiness, and interviews with our team.

Presenting the AWS Impact Accelerator Latino Cohort

New for this cohort, AWS accepted applications from startups headquartered in the U.S. but offering products and services to customers in Latin America. Almost half of the accepted companies are operating the region, including Argentina, Brazil, Colombia, Mexico, Venezuela, and Puerto Rico—this compounds the positive effect of the AWS Impact Accelerator to more people in more places.

Each of the selected startups was paired with mentors and technical experts that will advise them throughout the 8-week program. They will also receive up to $225,000 in cash and AWS Activate credits, curated training curriculum, introductions to AWS leaders and teams, networking opportunities with potential investors, and ongoing advisory support. Founders also have the opportunity to build their network and foster friendships and partnerships with their fellow program participants that will outlast the intensive 8 weeks.

United States & Puerto Rico startups

Alvva | Founder: Sergio Torres | Location: California
Industry: Financial Services

Alvva is an immigration platform built for the 21st century, offering immigrants to the U.S. a bilingual one-stop shop for filling out forms and getting financing for government fees. The company was built by founders who moved to the U.S. in pursuit of a better future, and who understand the barriers immigrants face when trying to file an immigration form.

DivySci Software | Founder: Ariana Abramson | Location: New York
Industry: Education

DivySci is a communication platform that provides real-time personalized feedback to adult learners, helping them develop practical digital communication skills for the workplace. Their robust speech technology platform covers diverse multi-cultural backgrounds to deliver objective measurements of communication signaling and behavior change algorithms for on-demand equitable learning.

Ease | Founder: Mario Amaro | Location: Texas
Industry: Healthcare

Ease is a fintech platform that offers healthcare professionals a comprehensive solution to start, grow, and manage their private practices. The platform features care delivery infrastructure, bookkeeping, and payroll services, as well as real-time payments, making it easier for practitioners to handle financial transactions and focus on delivering quality care. To date, they’ve helped build over 300 practices and are backed by top investors Slauson & Co. and Precursor Ventures.

EducUp | Founding team: Carlos Raul Garcia, Yusnier Viera, Yamel Barroso | Location: Florida
Industry: Education

EducUp is an artificial intelligence (AI)-powered educational platform that enables online educators to create gamified, engaging content for learners worldwide. With a rapidly growing community of 1.5 million students and a diverse course catalog, the company is transforming the online education space by making learning enjoyable, accessible, and effective for learners of all ages and backgrounds.

GamerSafer | Founder: Rodrigo Tamellini | Location: California
Industry: Gaming

GamerSafer is an innovative technology company that provides comprehensive security solutions for gaming platforms to protect their players from fraud and other crimes, as well as abuse and harm. With cutting-edge identity management software that uses computer vision and AI technologies, this cross-platform solution currently protects over 15 million players across 53 countries.

Lazo Fintech Inc | Founder: Juan Manuel Barrero | Location: Florida
Industry: Financial Services

Lazo is an all-in-one solution offering cost-effective legal, financial, and investor relations services for pre-seed, seed, and Series A startups. With a comprehensive ecosystem of tools— including a dashboard, data room, and Verified Customer stamp—Lazo enables founders to focus on product and traction, while ensuring they’re always venture capital (VC)-ready.

Leantime | Founders: Gloria Folaron & Marcel Folaron | Location: North Carolina
Industry: Developer Tools

Leantime is an open-source project management system that enables anyone to plan and execute projects using a combination of design thinking, lean, and agile best practices. With Leantime, even non-project managers can define project strategies, set goals, ideate solutions, plan timelines, and deliver on tasks without requiring any prior project management experience.

Monadd | Founder: Jessica Mendoza | Location: Florida
Industry: Financial Services

Monadd is a fintech and consumer tech company that simplifies and automates home bill management, empowering residents to take control of their bills and subscriptions through their AI-powered software. They work with property management, real estate, and relocation companies to ensure liability-free spend and services management for residents.

PilotoMail | Founders: Sofia Stolberg & Juan Carlos Stolberg | Location: Puerto Rico
Industry: Remote Work

The PilotoMail platform offers virtual mailboxes and mail management automation for remote workers and co-working operators. Its user-friendly interface helps users manage postal mail efficiently, while also enabling mailbox renters to access, track, and manage their mail and packages from anywhere in the world. It’s a compliant mail solution designed for the decentralized age of work.

Sign-Speak | Founder: Yamillet Payano | Location: New York
Industry: Hard Tech

Sign-Speak is an innovative technology company that provides automated American Sign Language recognition, transcription, and production to help Deaf and Hard of Hearing (D/HH) individuals communicate effectively in any situation, both in-person and online. Their solution uses animated signing avatars for impromptu interactions to create a more natural, dynamic communication experience for all.

STIGMA | Founder: Ariana Gibson | Location: Illinois
Industry: Wellness/Fitness

STIGMA is an award-winning mental health mobile app that provides anonymous one-to-one support in a low pressure atmosphere through text, audio, and video messages from people who share similar experiences. The app works with vetted mental health and wellness providers to offer members easy access to relevant resources and trusted mental health content.

Latin America (LATAM) startups

Cogniflow | Founder: Marcelo Martinez | Location: Colombia | US HQ: New York
Industry: Analytics, AI

Cogniflow is a no-code AI platform that allows entrepreneurs, scientists, and non-technical professionals to easily integrate AI into their daily tasks. Users can access pre-built AI models or create their own models using GPT-4 to solve a range of problems, from facial recognition to medical imaging analysis, increasing productivity and streamlining operations.

Deskfy | Founder: Victor Dellorto Toscano | Location: Brazil | US HQ: Delaware
Industry: Retail

Deskfy is a marketing portal that helps brands streamline their marketing operations and distribute promotional assets to their stores and branches, complete with insightful campaign reporting. It’s a leading Brand Management solution in the LATAM market, with global growth ambitions driven by successful partnerships with Audi, Levi’s, and Domino’s Pizza.

Fielder | Founders: Carolina Nanni & Jorge Villatoro | Location: Mexico | US HQ: Delaware
Industry: Marketplace

Fielder is an Enterprise software-as-a-service (SaaS) that digitizes, automates, and manages specialized technical services, including service orders documentation, asset and inventory management, and business intelligence for the future of work. They empower specialized technicians for better economic opportunities while optimizing costs and scaling flexibility for customers across industries such as information technology (IT), telecommunications, and more.

Kigüi | Founding team: Mauricio Kremer, Maximiliano Dicranian, Gonzalo Castro Peña | Location: Mexico
Industry: Food/Beverage, Sustainability

Kigüi is a money-saving grocery app that helps users find and buy discounted food close to its expiration date. Users can also get cash back by submitting their receipts and helping others find discounted products, creating a positive impact on the environment, community, and individual expenses. Kigüi is committed to both eliminating food waste and helping low-income households.

Kuentro | Founding team: Julio Pazos, Manuel Romero, Deivis Millan, and Hector Tamayo | Location: Venezuela
Industry: Sourcing/Recruiting

Kuentro is an affordable hiring platform that caters to middle- and working-class job seekers. The recruiting tool offers a space for small- and medium-sized enterprises (SMEs) to share job listings and make qualified hiring decisions—all without the need for a large human resources (HR) budget. The platform has a community of over 150k users, with 10.5k job offers and 360k applications.

Outtrip | Founders: Liliana Barck & Gonzalo Rico | Location: Argentina | US HQ: Delaware
Industry: Travel/Tourism

Outtrip is software for tourism adventure in Latin America. It helps small and medium tour operators streamline their operations through an easy-to-use booking system, customizable trip itineraries, and real-time availability updates. Outtrip helps operators save time, effort, and costs, while providing exceptional adventure experiences for their customers.

Panda Salud | Founders: Angela Cois & Guillermo Mogollán | Location: Mexico
Industry: Healthcare

Panda Salud is a one-stop-shop platform that connects healthcare small businesses with diagnostic test providers, online pharmacies, and medical equipment vendors to improve healthcare experiences and affordability. By fixing the fragmentation of the healthcare market in Latin America, Panda Salud empowers doctors and clinics to deliver better care.

Tienditapp | Founder: Luis Andrés Hernández | Location: Mexico
Industry: Freight

Tienditapp is an app that connects micro mom-and-pop stores with manufacturers and providers through a last-mile delivery model. Users can digitize supply chains, access fintech solutions, and analyze point-of-sale data to reduce logistics costs and develop new strategies at the point of sale. Tienditapp aims to serve 1,000 micro mom-and-pop stores in Mexico and more than 20,000 across Latin America by 2025.

TuCuota | Founding team: Juan Pablo del Peral, Fernando del Peral, Santiago del Peral, and Federico Isas | Location: Argentina
Industry: Financial Services

TuCuota is a fintech startup serving 100+ clients from diverse sectors. The company streamlines recurring payment collections for organizations in Latin America by consolidating multiple payment providers into one API. TuCuota’s mission is to provide a seamless, user-friendly platform that reduces operational costs and ensures secure payment experiences for customers.

Bessemer’s Jeff Epstein on the best advice to startup CFOs navigating uncertain times

Kate Radinovic — Thu, 27 Apr 2023 14:09:03 +0000

You are reading the second installation of our new thought leadership spotlight, “The Evolving Role of the Startup CFO.” This series features perspectives from prominent players in the startup ecosystem. Leading our second spotlight is Jeff Epstein, Operating Partner at Bessemer Venture Partners.

It’s no secret that market conditions are tough right now. But for startups, hunkering down isn’t an option: a startup has to build and grow in order to survive. So, what’s an ambitious business to do?

Rather than growth at any cost, “you need to grow in a rational and balanced way,” advises Jeff Epstein, Operating Partner at Bessemer Venture Partners (Bessemer). Over a long history that stretches back to the Carnegie era—and through plenty of trial and error—Bessemer has learned how to identify investments that can go the distance and build businesses that are fit to last. As the leader of Bessemer’s CFO Council, with plenty of direct CFO experience of his own, Jeff specializes in the kind of financial strategic and operational excellence that supports dynamic startups at many different stages of growth. We sat down with Jeff to get his best advice for CEOs kick-starting their company’s finance function and CFOs positioning their businesses for long-term success in uncertain times.

Get the right person at the right time

Startup businesses at different stages of growth will benefit from having different profiles in the financial leader seat. “At a company with about $5 million in revenue, the finance leader is very hands-on – both a player and a coach,” Jeff says. “Fundraising experience is not as important in this early stage, because the CEO usually leads fundraising. When a company reaches $50 million in revenue, the finance leader often fills three different functions simultaneously. They should have a keen, nuts-and-bolts accounting knowledge; they should be able networkers, ready to pound the Wall Street pavement and raise money from bankers and investors; and they should be eagle-eyed analysts, adept at forecasting, budgeting, and financial planning.”

Equally important to these skills is the working relationship between a CEO and a CFO. “Entrepreneurs are optimists,” Jeff says, “they believe things can be ‘bigger and better,’ and have an unreasonable confidence they can overcome all obstacles.” The board and investors, on the other hand, more often have a more balanced view. “At the best companies, the CFO is a counterbalance to the CEO, offering an alternative, more cautious point of view. Ultimately, the CEO makes the decisions.” Jeff says, “If there’s mutual respect between the CEO and the CFO, the CEO will make better decisions.”

Distinguish hopes from expectations

“The first rule of a CFO is: never run out of cash.” In today’s market, when capital is so expensive—and thus harder to come by—this caution is more important than ever, Jeff says. “It requires discipline that many companies didn’t need when they were growing 100% or 200% each year— they need to have it now.” At Bessemer, portfolio companies often aim for what Jeff calls the “Goldilocks Budget,” an aspirational plan that includes a contingency to balance the chance of success against the probability of falling short. It’s better to discuss probabilities openly with the company’s board, even welcoming disagreement about the best approach—so long as everyone gets on board with the ultimate decision.

Planning ahead and building in a buffer have always been sound strategies; now they matter more than ever. “The general framework hasn’t changed,” Jeff says, “For venture-backed companies, your cash runway, the number of months you have until you’re out of cash, is critical.” In 2021, companies raised venture capital in three months or less; today, it may take six months or more. In this market, the best CFOs try to extend their cash runway to two years by keeping costs low. They under-promise and over-deliver.

Don’t be afraid to experiment

Startup CFOs shouldn’t only hedge against risk. It’s equally important to remain innovative and nimble when it comes to the company’s day-to-day operations. Think about how to do more with less,” Jeff advises. “For instance, consider whether expenses like software and office space could be streamlined to cut costs and help the company work more efficiently. Companies are now closely measuring productivity,” he reflects. “For instance, do you need to pay New York or San Francisco wages to employees living in lower cost locations?

Big gains can also be found when companies scale up by improving their processes and automating —important evolutionary changes that companies might not have prioritized in an easier market environment. “Maybe you haven’t done that because you weren’t focused on process improvement as you were focused on growth,” Jeff suggests, “Now is the time to go back and improve all those processes.”

Build a team you can be proud of

In the end, piloting a company’s finances through choppy waters isn’t so different from smoother sailing. At least, the secret to a satisfied CFO is the same as ever: “the number one thing is being on a winning team.” And that still means scoring big, but home-runs might be harder to come by in a tough market. So, it’s more important than ever for CFOs to be proud of the teams that they build. “Not only your peers, and your CEO, and your board, but the people that you hired, and recruited, and trained,” Jeff emphasizes.

With skill-building and collaboration-enabling companies like Guild Education, Box, and more in the Bessemer portfolio, Jeff knows intimately the importance of assembling the right people and giving them the right resources and capacities for the job. A strong CFO can transform a somewhat experienced and partly trained team with less-than-ideal processes and systems into a well-oiled machine. Whatever the outcome in uncertain times, that’s something to be proud of.

In the next installment of our Evolving Role of the Startup CFO series, look out for some key tips from Chris Garber at Guild, a Bessemer portfolio company.

8 Highlights from AWS Startup Day Miami you want to experience

Chalaire Miller — Fri, 21 Apr 2023 13:46:19 +0000

Tech enthusiasts, engineers, startup founders, vendors, and more–from Miami and beyond–attended AWS Startup Day Miami for the kickoff of Miami Tech Month. Designed for startup founders and leaders, this full-day AWS event provided the startup community with the latest information on AWS services, solutions, and best practices. AWS Startup Day delivers education and networking opportunities to enable and accelerate startup innovation and growth, provides education on emerging trends, and inspires action through the personal experiences of startup founders, at no cost to attendees.

Whether you are a new startup looking to launch or an existing startup looking for ways to grow your business, you will learn something new from AWS Startup Day Miami, as well as the upcoming AWS Startup Day San Francisco on May 3rd – 4th.

The startup-focused agenda and lineup marked a great start to Miami Tech Month, which is an annual event that celebrates and showcases the thriving technology ecosystem in Miami. It usually takes place during the month of November and features various activities such as conferences, meetups, hackathons, workshops, and networking events. The goal of Miami Tech Month is to promote collaboration, innovation, and entrepreneurship within the tech community in Miami and to highlight the city’s potential as a hub for technology startups and talent. Kicking things off with AWS Startup Day Miami just makes sense.

Here are some of our favorite moments and takeaways from the AWS Startup Day Miami event.

The opening montage really set the tone and created excitement by capturing the attention of the audience and reinforcing AWS Startups message of “customer first” and “startup focused.” Experience the full opening montage for inspiration and knowledge to prove what’s possible.
Alvaro Echeverria, GM of LATAM Startups, kicked things off with an insider talk on Amazon’s Culture of Innovation, dropping several gems and takeaways for decision-makers in the startup ecosystem.

Alvaro Echeverria shares his knowledge with the audience.

Check out the video of Alvaro’s talk.
The keynote was delivered by Erick Gavin, the Executive Director of Venture Miami, an economic development office of the City of Miami tasked with facilitating the growth of the technology and innovation ecosystem. Gavin encouraged attendees to utilize AWS solutions to scale and grow, and he doubled down on his office’s commitment to continue providing tech startups opportunities and access to various resources and tools.

During his keynote speech, Erick Gavin of Venture Miami shares advice and support with Startups: “Take advantage of diving into AWS events and resources.”
The AWS Generative AI Accelerator was announced – a 10-week program designed to help early-stage startups using Generative AI solve big challenges to scale and grow. This gave attendees the opportunity to engage and ask questions of AWS experts in person about the program and the application process. Stay tuned to the AWS Startup social platforms, as the selected Generative AI Accelerator participants will be announced soon.
The panel discussions like “Raising Money in Miami” were lively, well-attended, and the panelists were very intentional with their advice. That panel included Andres Barreto of Techstars, David Blumberg of Blumberg Capital, Alexandra W of Clerisy Capital, Joshua Siegel of AcronymVC, and founder Yasmine Morrison.

The Raising Money In Miami panel discussion.

Learning new ways to raise money to fund your startup is never a bad idea. Watch part of the panel discussion here.
Hearing from AWS customers and partners about how much the AWS Startup Day Miami event means to them. Most of the sentiments were made spontaneously, like by Macarena Bravo of Bravo Consulting, who spoke about how much she loves being an AWS cloud customer and attending AWS events and trainings. This led to the AWS team recording a series of customer videos.

Macarena Bravo of Bravo Consulting attends Startup Day Miami.

Hear what attendees say about Startup Day Miami.
Networking happy hour. Many of the speakers were in attendance for this portion of the event. In any industry it is important to build relationships, but more so in the tech startup world. It is important to continue to learn, and stay connected with the larger tech community as new challenges and changes come up every day. Building that community helps with easing the burden of going it alone.

People enjoying everyone’s favorite hour of the day.
Last but not least, come meet the AWS Startups Team and let us answer your questions. In learning more about AWS services for startups, it’s great to have that friendly face or cloud expert to discuss potential opportunities and the wide range of AWS services that fit your startup needs.

***Ready to attend an AWS Startup Day and* accelerate your startup? Check out the upcoming AWS Startup Day San Francisco *on May 3rd – 4th.***

Stay in-the-know for all things startups:

AWS Startups Twitter for the timely and behind-the-scenes content
AWS Startups LinkedIn for the latest insights
AWS Startups Facebook for the biggest news
AWS Instagram for the motivational and inspirational

Andreessen Horowitz cofounder Ben Horowitz on the cloud, weathering economic turmoil, and building lasting relationships

Megan Dubinsky — Mon, 17 Apr 2023 16:56:44 +0000

Few people have seen the technology sector transform like Ben Horowitz. The Andreessen Horowitz (a16z) cofounder, entrepreneur, and New York Times-bestselling author has spent decades fostering innovation and building strong business relationships—and has plenty of wisdom to share.

Horowitz spoke with Ruba Borno, Amazon Web Services (AWS) Vice President of Worldwide Channels and Alliances, at the AWS ‘Startups in the Stadium’ event at re:Invent this past December. Here are six major insights from their discussion—on trust, venture capital, the cloud’s impact, and more.

Ben Horowitz and Ruba Borno at the AWS Startups in the Stadium event at re:Invent.

Today’s VC industry empowers founders

Since a16z began in 2009, Horowitz has witnessed – and spearheaded – several major shifts shaping the venture capital (VC) industry. For one thing, he explained, it’s more likely today that founders will lead their businesses through all stages of growth, instead of leading from “zero to one” and then ceding the role to a professional chief executive officer (CEO).

That shift is due partially to the philosophy that a16z brought to the VC world. “Our idea was … what if we build a firm that enables the founder to be a CEO?” Horowitz explained. After all, a leader who truly understands the company could shepherd the business through another winning project—a secondary product, for instance.

He cited Amazon Web Services itself as an example. “It’s another [project] that was done much later. And we all know if Jeff [Bezos] wasn’t the CEO, it would never have happened. You can’t bring in somebody who doesn’t really understand the company to go from zero to one.”

Cash is (and remains) king

As much of the world heads into economic uncertainty, Horowitz warned startups against committing what he views as business’ one “unforgivable sin”: running out of cash. In a downturn, this becomes particularly difficult, as consumers tend to cut newer companies from their budgets before tried-and-true stalwarts.

Still, Horowitz said, it should remain the priority. “You can screw up absolutely everything: the product, the go-to-market, everything. But if you still have money, you can fix it … if you run out of money, it doesn’t matter how many good things you do. You’re dead.”

The cloud drives innovation

Many of a16z’s portfolio companies build on AWS, which Horowitz sees as a means to scale, innovate, and streamline day-to-day operations.

“The cloud deployment model is just such a massive advance in every person’s ability to build great software, solve big important problems, and focus on higher level issues,” he said.

He pointed to the difference between bygone tech companies like Netscape, which sold its Navigator browser as a physical, boxed product, and newer players like Snap, which created and distributed its product with just, “four guys, a laptop, AWS, boom.”

“[AWS] has been a big change,” he said. “Which is why there are so many more startups now.”

In addition to boosting innovation, AWS helps smooth out internal processes, he said. Old-school software was difficult to install and implement—plus, training teams on new software (and updates) could be arduous. Integrating new software was so difficult, Horowitz explained, that a company’s chief information officer (CIO) was the point-of-contact for sales. But when the cloud came along, it became much easier, and software companies could sell to whichever department had the actual business need.

“You don’t need training,” Horowitz said. “The software is usable enough and away you go.”

Three ways to succeed with AWS

Horowitz had three pieces of advice for businesses interested in working with AWS.

First, he said, it’s imperative to get into the AWS Marketplace. “It works,” he said. “You will get sales, it’ll help you.”

Next, he emphasized the importance of aligning your sales teams with AWS sales. “Your regional salespeople need to find their AWS sales counterpart, who want to help you. Your teams should find them … call them up … it’s maybe your best lead-gen opportunity.”

Finally, he recommends all businesses go through the AWS technical review process. “The reason you should go through that is: If you’re in the AWS field, how do you know if this software is any good?,” he said. The way to find out: approval from a trusted party like AWS.

Building trust within partnerships

Horowitz shared his organization’s culture values trust and communication. Here’s three ways a16z builds trust across partnerships.

First, he said, “we have one culture for how we treat each other and how we treat people outside [the organization]. We don’t change it. It’s not us and them.”

Next, his word—and the word of his team—is a bond. “You don’t need a contract with us,” he said. “If we say it, we’re going to do it. We tell our employees to be careful about what they say because they’ll bind the company.” (In fact, a16z employees must sign a culture document affirming that they understand this.)

Finally, he emphasized that he does not believe in transactional relationships. “We take the long view of relationships because we’re in the relationship business,” he explained. “We’re either in business with you for the long term or we’re not.”

Creating the future you want

Horowitz has held a variety of roles throughout his career. As Borno put it, he’s been “a product manager, CEO of a startup to GM at HP, and … founder of one of the most successful VC firms in the world.”

Horowitz credited this path not from a desire to reinvent himself, but rather a willingness to step up and do what’s needed. He recalled that in the early days of Netscape, where he worked as a product manager, his future partner (and cofounder of Netscape) Marc Andreessen pointed out that if they didn’t take steps to bolster the early internet, there was no guarantee that someone else would take on that work. So, they did and that proactiveness paid off: It helped usher in SSL, JavaScript, and more.

Horowitz has carried that attitude with him in his subsequent endeavors: The priority is always a positive impact on the customer.

“What do I need to step up and do?” Horowitz said. “What am I capable of doing that’s going to make a difference?”

See the full interview here.

Evolutionary architectures series, part 3

Aayzed Tanweer — Wed, 05 Apr 2023 15:43:50 +0000

Building a scalable tech stack is an iterative process for startups.

“To the moon ?”

“Evolutionary Architectures” is a four-part blog series that shows how solution designs and decisions evolve as companies go through the different stages of the startups lifecycle. In this series, we follow the aptly named Example Startup whose idea is to create a “fantasy stock market” application, similar to fantasy sports leagues. They envision holding four “tournaments” over the course of a year.

The second blog described how the startup started evolving their technical solutions while the founders were getting ready for fund raising. In part 3, we will see how Example Startup further progresses in maturing their tech stack and positioning themselves well for scale.

Scaling efficiently by transitioning to a microservices architecture

The fantasy stock trading team is growing and new components and solutions are being built. As the technical portfolio expands, certain cracks begin to appear that required the team’s attention.

“Old habits die hard,” and the team begins to see how this can cause problems for their startup’s growth: The aggressive timelines and the enthusiasm to get more done with less is leading to increasing technical debt. One aspect of this technical debt is a gradual proliferation of monoliths, as opposed to the microservice architecture that the team had initially decided upon. Monolith concerns such as scalability and performance bottlenecks begin to show during testing and the introduction of new features. Luckily, the team quickly recognizes the challenges this monolithic approach poses to the optimal scaling of workloads. They decide to take a step back and reevaluate their development practices. One of the developers remembers that the AWS solutions architect (SA) had anticipated some of these problems in an earlier conversation. The Example Startup team schedules a call with AWS to get some help.

Breaking down monoliths and transitioning into a microservices-based paradigm is a broad topic so the AWS SA recommends an App Modernization Immersion Day for the team at Example Startup. The immersion day uses a related workshop as a backdrop, with a focus on workloads relevant to startups. The event is attended by nearly all developers at the company and ends up being a game-changer. Over the course of a single day, the team is able to learn how to properly define, design, and implement microservices. They also learn about charting a gradual migration path from a monolith application to a set of microservices without having to redo everything at once. The team is glad to catch their mistakes early on and learn some best practices that will help them going forward. The solutions architect also shares an AWS whitepaper focused on modernization strategies that can fill in any knowledge gaps on the Example Startup team.

The experience with app modernization provides so much value to Example Startup that the team decides to apply the same approach of leveraging existing best practices for different functional areas going forward. The engineers and product manager schedule a call to share their roadmap for the remainder of the year with AWS, in an effort to avoid duplicative work. Example Startup already signed a mutual non-disclosure agreement (MNDA) with AWS and there was a productive free flow of ideas across both sides during this conversation, as well as some great news: It turns out that a feature Example Startup was considering building out themselves is already on AWS’ roadmap for the next quarter, and this frees up a chunk of engineering time for the team.

The next topic on Example Startup’s list of areas to improve relates to Infrastructure as Code (IaC), continuous integration and continuous delivery (CI/CD), and automated testing. Two newly hired developer operations (DevOps) folks aren’t satisfied with many of the current operational mechanisms at the startup, especially things like building and testing environments, as well as managing code artifacts. A growing team at Example Startup means that more people have access to these sensitive processes, thereby introducing unnecessary risk. The two new team members already have some experience with Terraform as their approach to IaC. They are happy to learn that AWS is well supported by Terraform, and to discover other tools like AWS CloudFormation and AWS CDK in case an alternative is needed. However, they still need some help with their CI/CD setup. Their attempts insofar lack cohesion and it proves difficult to make their build tool work well with their deployment tool. Additionally, they are still looking for a suitable approach to manage their container images. The AWS team recommends looking at AWS CodePipeline because it meets their needs for integrating a build and a deployment tool seamlessly and also includes automated testing, all paired with support for various environments. Using CodePipeline allows integration with solutions that weren’t necessarily built natively on AWS, as well as robust support for other tools such as AWS CodeBuild, AWS CodeDeploy and third-party tooling. Implementing CodePipeline allows Example Startup to check off another big item of their list.

With the team well on its path to a proper implementation of microservices, they feel empowered to work on some of the other complex challenges that remain outstanding. For one, the presence of multiple services operating independently naturally brings up the question of communication across these services. There is a big question mark around whether every cross-service call should be synchronous or asynchronous in communication, in addition to how the team can begin adopting best-practices patterns such as publish/subscribe (PubSub) messaging. The team understands broadly that adopting an event-driven architecture would be beneficial, especially with the move away from monoliths, but they are a little overwhelmed with the endless array of AWS services related to that architecture, including but not limited to Amazon EventBridge, Amazon Simple Queue Service (Amazon SQS), Amazon Simple Notification Service (Amazon SNS), and Amazon Managed Streaming for Apache Kafka (Amazon MSK). This time around, the team is able to find some resources themselves as a great starting point such as some very useful workshops and blogs on the topic. The “event driven” paradigm is slowly becoming another tool in the team’s toolbox.

Developing a stronger security strategy

Security continued being top of mind for our startup and tools like the AWS Startup Security Baseline (AWS SSB) help them to get started. Unfortunately, you can never have too much security. The initial implementation of AWS WAF was a good start, but the team needs to start thinking more proactively about prevention, detection, and remediation. They begin upskilling themselves on the many AWS services focused on security that can help them implement a strong security strategy.

The growing team and the involvement of partners makes access control, permissions, and governance other topics requiring an increasing amount of focus. The team is trying to implement best practices such as the principle of least-privilege when applying permissions. At a minimum, they want to move the production workloads into their own, separate accounts. As the team adopts these best practices, they see the increase in operational complexity due to the added layers of management and permissions they are now having to deal with. It becomes rapidly obvious that they need a mechanized approach to account structure. Someone mentions AWS Organizations,which seems like a step in the right direction so they reach out to their trusty AWS SA for a chat. The SA shares some relevant advice, like looking at AWS Control Tower as an easier approach to managing multiple accounts and AWS Organizations. Since this is the first of many steps towards achieving a robust multi-account strategy, the AWS SA also shared with the team the “Transitioning to multiple AWS accounts” prescriptive guidance. This guide includes best practices around account migration, user management, networking, security, and architecture when moving to a multiple accounts setup.

Optimizing workloads for performance

The team is tackling some foundational pieces so the startup will be well poised to grow at the right pace. A few major items are crossed off the list and others have action plans in place. The developers are doing as much as they can to optimize their workloads for performance, but have identified some opportunities for further improvement that go beyond code, such as edge caching with Amazon CloudFront, caching on an application level with Amazon ElastiCache and Database caching. The team is increasingly growing reliant on AWS Managed Services to give them the functionality they need while keeping the associated operational complexity at a minimum. Another managed service that some of the developers discover and find surprisingly easy to use is AWS Batch. The initial feed processing approach with AWS Lambda is starting to hit its limits due to the exponential increase in the volume of data that needs to be processed. After some experimentation, developers are able to chart a path to using AWS Batch that allows them to keep growing with relatively little increase in operational burden and while keeping costs low.

The updated AWS architecture diagram for Example Startup

Proving their startup’s value proposition

All this good work at Example Startup does not go unnoticed. Building in an agile-yet-sustainable manner without reliance on short-term workarounds shows that the company is thinking about the long term, displays maturity, and has the capability to deliver. These traits along–with an innovative solution and a good product market fit—are at the core of the company’s value proposition. The founders successfully convey their company’s value to couple of different venture capital firms and close their first Series A funding round. Example Startup is on its way to the moon.

Check out the first blog and second blog in the Evolutionary Architectures series.

AWS launches global generative AI accelerator for startups

Howard Wright — Tue, 04 Apr 2023 10:27:16 +0000

Building on the exciting new developments around generative AI, we are happy to launch the AWS Generative AI Accelerator, a 10-week program designed to take the most promising generative AI startups around the globe to the next level.

Generative AI is an incredibly exciting field that has the potential to revolutionize many industries. Tech startups will play an important role in bringing this technology into the mainstream. Take the film industry, for example. The visual effects that make heart-pumping action scenes, amazing super heroes, and enthralling new worlds take months of tedious work by hundreds of artists to come to fruition—adding up numerous hours in post-production time and using millions of dollars in production budget. Because visual effects are so expensive and time-consuming, films with more modest resources struggle to implement high-quality visual effects to bring forth their vision.

Runway, one of our startup customers, is helping artists revolutionize filmmaking with their AI Magic Tools—some of which have been used in the award-winning hit “Everything, Everywhere, All At Once.” Their text-guided generative diffusion models are unlocking powerful new multi-modal creation and editing solutions for artists. With Runway, the difficult tasks of composition, stylization, inpainting, motion tracking, and other processes are made easier and quicker for creators, allowing them to focus on more idea concepts and to deliver faster iterations. These tools also cut down production costs and lower the barrier for filmmakers—professionals and amateurs alike—to push the boundaries of movie making and let their imagination run free.

In addition to its creative potential, generative AI has numerous practical applications. It can be used in healthcare to create personalized treatment plans or to better analyze medical images; in finance, it can generate smarter analysis and draw insights; in tech, it can write code and reduce human-error bugs; in manufacturing, it can design new products and optimize production processes.

Here at AWS, we believe the startup community will be the driving force moving these innovations forward. The AWS Generative AI Accelerator is designed to act as catalyst, helping some of the most promising companies in this space to take their ideas off the ground. With a program tailored to meet the needs of generative AI startups, the AWS Generative AI Accelerator will provide access to impactful AI models and tools, customized go-to-market strategies, machine learning stack optimization, and more. Selected startups will also have access to networking opportunities with industry luminaries, potential investors, and customers. In addition, the selected startups will receive up to $300,000 in AWS credits to build their products and services on our tech stack, as well as dedicated business and technical mentors matched based on industry vertical, market, and stage.

To fully benefit from the program, startups should have a minimal viable product (MVP) already developed, some traction with customers, and be working to enhance their product value proposition in order to scale. Although the program is open to all startups, those already building on AWS will receive the most benefit from the accelerator’s dedicated AWS Solutions Architect team, who will support every step of their product development. The program is open to companies around the globe, with no limitations around use case—we want to empower companies applying generative AI to solutions from legal and marketing, to software engineering, green energy, and life sciences, including drug discovery.

Here’s what some of our investor partners are saying about the program:

“AWS Accelerator offers a compelling program for founders building in the Generative AI space, and I look forward to meeting the cohort when they’re announced,” said Sonya Huang, Partner at Sequoia Capital.

“With so much activity and opportunity today in the world of AI-driven startups, it is a great time for AWS to launch this Generative AI Accelerator. I am looking forward to seeing which companies are selected for the program and how AWS will partner with them to turbocharge their growth,” said Rob Toews, Partner at Radical Ventures.

We are excited to meet the founders who are applying generative AI to solve some of our society’s most riveting challenges.

Applications for the AWS Generative AI Accelerator will be open today through April 17, 2023. To learn more and apply, please visit AWS Generative AI Accelerator.

How machine learning helps Fraud.net to build a modern app on AWS to combat financial fraud

Megan Crowley — Mon, 03 Apr 2023 12:03:51 +0000

Startups know firsthand how better technology can improve the quality of life: From AI/ML allowing scientists to better predict patient health outcomes, to cloud computing driving life-saving innovation, and modern apps enhancing accessibility.

With better technology also comes the opportunity for criminals to commit more advanced levels of crime. Fraud, especially, is occurring with greater technical sophistication as society transitions into a digital-first world. Fraud and cybercrime are also growing at significant rates, and now cost businesses around the world over $6 trillion per year or an average of 5% of their revenues.

To outpace and outsmart the technology criminals use to commit fraud, former bankers Whitney Anderson and Cathy Ross founded Fraud.net, a modern fraud and compliance platform, in 2016. Fraud.net offers customers in the banking and fintech industries across the globe a serverless modern application that uses artificial intelligence and machine learning to rapidly identify fraud, leading to more efficient operations and higher customer satisfaction.

Providing a modern solution to an evolving problem

Whitney Anderson, co-founder and chief executive officer (CEO) of Fraud.net

As is the case with many successful startups, Fraud.net encountered a challenge and saw the opportunity to build a solution that helps themselves and other companies to overcome it.

“We were our own use case,” explains Whitney. While operating companies in the digital commerce and payments world, “One of the greatest frustrations was experiencing multi-percent fraud rates, and payment processors not giving us access to the information we needed to solve the fraud.”

To solve a problem that caused harm to companies and customers alike, he explains, “We started pooling together other players in the digital world: payment facilitators, merchants, and other ecosystem participants.”

“By sharing secure and anonymized data, we were able to reduce fraud by more than by 66%. It was simple, immediate, and intuitive.”

One major finding was that the same people used the same technology-based methods to defraud multiple companies. “It’s really difficult to fight fraud by yourself,” says Whitney, “Sharing data in a safe and secure way let us understand a lot more about the bad actors and separate them with the goal of really delighting the 99% good customers.”

Sharing an enormous amount of information within their cross-industry consortium meant that Fraud.net needed a rapid, scalable solution to unify their data and create real-time actionable insights.

Fraud.net chose to go all-in on Amazon Web Services (AWS).

Building an event-driven architecture on AWS

As a cloud-native modern app, Fraud.net uses an event-driven architecture that uses serverless components. Event-driven architecture makes it more efficient for startups to develop modern apps because they scale up to address events and scale down when no events occur. This can result in saving the startup resources and costs, which is critical as startups go to market. One benefit of Fraud.net’s event-driven architecture is the scalability and speed with which their developers are able to bring products to market.

Fraud.net’s AWS solutions include EC2 and Lambda for compute, S3 for highly scalable object storage, and DynamoDB as their noSQL serverless database.

Together, these solutions help them to unify and analyze three levels of data: customer-level data, institution-level data, and cross-institution data.

“Because of AWS’ serverless technologies and other incredible innovations, we’ve been able to unify data for fraud prevention, anti-money laundering, and compliance functions,” says Whitney.

Events from the Fraud.net platform arrive through a Fraud.net API that is managed by Amazon API Gateway. When the events arrive, they trigger an AWS Lambda function to process records from Amazon DynamoDB.

“Lambda functions have been a game-changer for us. We ask thousands of questions for each application or transaction submitted to us for risk assessment, based on different scenarios and risk profiles. All of those would have needed to be done in our own data center with tons and tons of servers,” says Whitney. “Instead, Lambda and its serverless capacity help us answer those questions in milliseconds, and helps us achieve decision accuracy upwards of 99.9%. It’s hugely efficient and cost-effective technology for us and our clients.”

Fraud.net also uses Amazon Kinesis to process and analyze streaming data in real-time to give customers results based on the latest and most comprehensive data. Amazon Redshift is their data warehouse, which they use to conduct data analytics on incoming events, transactions, and more.

Per Whitney, “AWS helps us process thousands of transactions per second, at a scale that was virtually impossible three or four years ago.”

Going serverless for scale and speed

Whitney credits AWS serverless technology as a critical component in Fraud.net’s mission to make every digital transaction safe. “In the past, providing a unified suite of microservices to fight fraud is something that hadn’t been done, or certainly hadn’t been done effectively,” explains Whitney. “With some of the older siloed databases, it wasn’t even possible to do.”

“Serverless is also unbelievably quick and easy, relative to the old days with on-premise software when a bank could expect it to take six months to a year to integrate a system,” says Whitney. Fraud.net accomplishes most of their customer onboarding with a simple set of no-code tools that leverage a suite of APIs to onboard a bank or fintech within 30 days, including the planning and training time.

“Because it’s so cost-effective, we’re 99% serverless,” says Whitney.

Fraud.net offers one of their serverless products, Transaction AI—a transaction monitoring, fraud prevention, and revenue enhancement platform–on the AWS Marketplace.

Gaining actionable insights using machine learning

Fraud.net uses Amazon SageMaker to create, train, and deploy the machine learning models that provide their customers with an average of an 80% reduction in fraud cases, a 92% reduction in false positives, and a 30% increase in approvals for good customers that were erroneously flagged as high-risk.

Machine learning allows Fraud.net to provide banks and fintechs with answers in under a second that otherwise may have taken employees hours of time-consuming tasks, such as manually cross-checking client information. Whitney explains, “AWS technology as a baseline, with Fraud.net’s software layer on top, enables teams to be much more efficient and spend their time more wisely.”

“The underlying technology, along with Amazon’s pricing, enables us to ask about 20,000 questions about identities and behaviors every time we receive a new account application or transaction,” explains Whitney. “All of that gets handed off to machine learning. We now routinely build clients custom ML risk models, with several hundred million features as inputs, because AWS has made it so relatively inexpensive to do.”

Teaming up with AWS to provide value to their customers

Cathy Ross shares her insights at an AWS event.

Alongside the AWS technology that Fraud.net uses to give their customers rapid and accurate tools to fight fraud, they also work with AWS to optimize customer costs. Whitney explains, “Our customer’s average return on investment (ROI) using Fraud.net is over 700%. That’s largely due to AWS’ efficiencies in cost structure. We leverage that and offer an incredible value to any company that uses Fraud.net.”

Fraud.net also collaborates with AWS teams in retail payments, financial crime, and other teams to give their clients a safe and effective onboarding experience. “We get a lot of support from various AWS teams,” says Whitney. “This is often a client’s first interaction with the cloud environment. Some of our big financial services clients come from on-premises environments, and they specifically come to us because we prove out a super-strong ROI from using their first cloud-based project.”

Looking to the future of fighting fraud

As a global fraud prevention management system, Fraud.net is, “all about scale at this point,” says Whitney. With clients ranging from top-tier financial institutions all the way down to early-stage fintech startups, and across industries such as financial, e-commerce, travel, and more, Fraud.net’s goal is to be the preeminent fraud and risk management layer for all digital enterprises.

For other founders looking to build a successful startup, Whitney advises three things that make a good entrepreneur:

Know an industry really well, see the gaps, and envision a better future for that industry.
Be a problem solver who gets excited about the prospect of fixing the problems that you see.
Have a deep reserve of energy and enthusiasm to get you through the good times and the bad.

For fintech startups in particular, Whitney advises that the upcoming FedNow Service launch in 2023 is likely to, “present a huge new set of risks and a need for risk to be solved immediately.” The FedNow Service is a real-time payments network that will allow money to transfer in seconds instead of in days.

With this advance in payments technology. Whitney expects to see an enormous amount of beneficial innovation on AWS and in the fintech world as technology ramps up to outpace fraudsters.

“It returns back to simple trust enablement,” he explains. “For banks and companies, it’s about restoring trust in your relationships with customers thousands of miles away that you’ll never meet.”

Curious about how AWS can help kick start your startup? Join AWS Activate to build and scale your startup with the right resources at the right time.

Check out more Fintech startups building and scaling on AWS ?:

How InsightFinder uses AWS solutions to build an AI-driven predictive observability platform

Helen Gu — Thu, 30 Mar 2023 13:53:16 +0000

Holding onto your time—to be a spouse, to be a community member, to be an individual—is difficult, particularly in the world of startups. InsightFinder, an artificial intelligence (AI) startup that uses machine learning (ML) to help customers prevent outages in their cloud infrastructure, is on a mission to change that.

Founded by Helen Gu in 2016, InsightFinder uses unsupervised machine learning to make cloud infrastructure more reliable. The company’s AI-driven predictive observability platform helps companies to predict business-impacting incidents as well as pinpoint the root cause of impending incidents to avoid business loss and brand damage.

Helen Gu, founder and CEO of InsightFinder

Helen says, “IT outages have a huge impact on everybody’s life. InsightFinder’s mission is to help everybody have a more reliable IT system.”

Fewer outages allow people more time to focus on doing what’s most important in their lives and for their businesses.

Alongside being the founder and chief executive officer (CEO) of InsightFinder, Helen is a professor at North Carolina State University and a distributed systems cloud computing expert whose work spans 20 years. “InsightFinder is created out of over 15 years of research work sponsored by National Science Foundation and industry partners,” she explains. “From day one, I’ve been very passionate about this field because I think it will affect a lot of people.”

Building InsightFinder’s solution with AWS

Amazon Web Services (AWS) played an integral role in InsightFinder’s development on the cloud. “When we first started, in the early days, we were looking for something that was easy to use,” Helen says.

“AWS has a very nice program, AWS Activate, that gives a lot of credits to startup companies. We got quite a lot of credits that helped us to bootstrap our development. That played a critical role for us.”

It’s also thanks in part to AWS that InsightFinder has been able to build the high-performance Unified Intelligence Engine that fuels its success. The company leverages AWS solutions according to their needs, whether for CPU-intensive processes or I/O-intensive processes. “A lot of AI tech companies think you need to invest heavily in hardware resources,” says Helen. Through AWS, “We can actually build a high-performance engine, and with reasonable cost.”

By 2020, InsightFinder saw the problems that their customers faced were evolving, mostly due to the Covid-19 pandemic. One startup, Apprendis, faced a problem of scale as the number of students and teachers using their platform for science education, InqITS, grew quickly. The company is an active Amazon CloudWatch user, but lacked the internal infrastructure to filter the alerts sent their way. By connecting the InsightFinder engine with the CloudWatch data, the company could receive essential insights quickly and easily. “It’s very simple,” says Helen. “A few clicks, and they can actually start to use the data and the predictions, and get the root cause analysis from AWS CloudWatch data through the InsightFinder engine.” Using CloudWatch data, InsightFinder caught hard-to-find software bugs and performance issues before end users noticed, which allowed the Apprendis team to scale their system without hiring DevOps engineers and to ensure seamless usage of its platform.

Now, InsightFinder is looking to the AWS Partner Network to push the company to the next level.

“Being a fast-growing tech startup company, we don’t want to hire a large sales force to directly sell to a lot of customers,” Helen says. “AWS Partner Network is going to be an important go to market motion for us. It’s a more productive, effective, mutual, beneficial path for us.”

Looking to the future of AI

As fears that AI could impact job opportunities grow, Helen is clear that she doesn’t believe the goal of InsightFinder, or AI as a whole, is to replace people: “There’s no way we can have enough people, particularly skilled people, sitting there looking at all those charts for each machine and predicting future incidents.” Instead, Helen sees AI as a tool to augment human skill; that the technology should focus on things that humans cannot do. The algorithms that Helen works with fill the gaps that humans cannot, performing 24/7 work that is suitable for a machine, not a human being.

“The most precious thing in the world is time,” says Gu. “Our impact is that we want to give time back to people to do things they enjoy, rather than fix IT outages in the middle of the night.”

Curious about how AWS can help kick start your startup? Join AWS Activate to build and scale your startup with the right resources at the right time.

Check out more AIML startups building and scaling on AWS ?:

Meet three women who make startup journeys smoother

Bonnie McClure — Wed, 29 Mar 2023 12:13:48 +0000

To celebrate International Women’s Day and Women’s History Month, we’re featuring posts throughout the month that highlight women in technology who are building and creating. Above all, these women are inspiring, empowering, and encouraging everyone in technology—especially women and girls— to prove what’s possible. Three of our Solutions Architects tell us how they got here, the coolest milestones they’ve helped businesses achieve, and the advice they have for other women who want to make it in tech.

Across AWS for Startups, women are working to make the startup space more equitable. Solutions Architects are a big part of that initiative, since they act as strategic advisors who make it easier for startups of every background to thrive. By guiding enterprising founders toward the resources and building blocks that will help them succeed, Solutions Architects break down barriers to entry and welcome startups into a trusted global ecosystem.

There’s no one-size-fits-all solution to the real-world challenges startups face, so AWS Solutions Architects get to flex both their creative and technical muscles as they design the personalized fixes that will help each startup thrive.

This Women’s History Month, we chatted with three AWS Solutions Architects who are helping startups of all backgrounds accelerate their businesses. Skye Hart is based in Denver, Colorado, and comes to the team with a background in data engineering. Jamila Jamilova uses her expertise as an economist to help startups working in the UK and Ireland. Aleena Yunus is based in Munich and specializes in analytics as she helps B2B-engaged startups achieve their goals.

Aleena Yunus Skye Hart Jamila Jamilova

How would you describe being a Solutions Architect?

Aleena: It can be anything from speaking at conferences to writing long emails. I act like a strategic advisor with regard to a startup’s roadmap and everything around infrastructure and architecting. We whiteboard together, or I sometimes help them build architectures, create a minimum viable product, or a proof of concept.

Skye: I like to think of myself as an extension of the startups I work with.

Jamila: I’m their trusted technical advisor, but I’m always looking at their core business and trying to understand, how can I help them optimize costs?

What are some of your favorite parts of your job?

Jamila: The diverse landscape. I am working with business decision-makers and with technical people, and this gives me a big picture of how the startup operates.

Skye: Every morning I wake up and I’m either putting on a lab coat and talking to a life sciences organization that’s changing the game for cancer research or I’m putting on my little construction hat and trying to work with mechanical engineers to develop IoT sensors.

Aleena: Creating usable content. I’m really passionate about sustainability and created a set of best practices. Ever since then, I’ve also been in touch with other customers who want to talk about sustainability.

You are women working in a startup world where there are still too many barriers to entry for people who are not wealthy, white, and male. What were some of the things that led you to this career despite those barriers?

Skye: In my first tech job [not as a Solutions Architect], I walked into a room and there were 35 guys and two women. I said, ‘This is just going to be that way.’ You have to work on overcoming imposter syndrome and being confident in yourself.

Aleena: It’s important to have role models. My older sister is in tech, and it was really important to see that, okay, there is disparity, but there are people who are making it.

Jamila: If you are in a company where you feel belongingness and where you feel that you are respected for who you are, for the skills you bring to the table, you don’t see any barriers. I feel like I am kind of a piece of a puzzle in a big picture. I can bring in my skill set, my experience, my worldview.

Jamila stands in front of the famous Jeff Bezos quote: “Work hard, have fun, make history.”

What are some of the times you’ve been able to help a startup customer achieve new goals?

Aleena: We helped one of my customers build a data pipeline from scratch in six weeks when they weren’t doing anything with the data they were storing. We wanted them to understand they could enable their sales team to get insights on it.

Jamila: I had a customer about to go on a fundraising round, but they desperately needed to show that they could bring the cost down, keep the same quality, and serve their customers as expected. My team worked to identify their architecture and the services they needed to optimize costs.

Skye: We had a healthcare company outgrowing their current environment. We had a leadership offsite with them and then kicked AWS into gear, and by developing relationships with business development on their go-to-market strategies, we organized immersion days for machine learning and security to get them hands-on training for up to 60 different engineers in multiple different cities.

What advice do you have for startups across the board?

Skye: Don’t be afraid to ask for help. AWS for Startups is an ecosystem. It lives and breathes for startups everywhere. Ask us what you need, what you want, what you dream about, and we’ll align the right things.

Jamila: Please do thorough research before you actually invest or get credits to run your business. Search for which cloud provider is best for your startup, and you will see lots of forums and resources. You will discover a whole world where you will be able to book free tech and business consultations.

Aleena: Picking the right tool for the right job is really important. If you don’t use the right building block, it’s not going to fit. Do your research about what is best for your use case.

Do you have any advice for students or people early in their careers who are interested in the startup landscape?

Aleena: Don’t get intimidated. If you have an interest, just go for it. What’s the worst that’s going to happen? You’re going to try and you’re not going to achieve it. It’s better than not trying at all.

Skye: Whatever industry you’re interested in, consider yourself a futurist. Ask yourself, ‘What is next?’ I see doctors thinking of ways that we can make telehealth easier and lawyers thinking of new ways that they make casework easier. Whatever you’re interested in, don’t think about what that position means right now, think about what that position’s going to mean in five years, 10 years, 20 years, and how tech can get you there.

Jamila: If someone tells you that you cannot be something, you can and you will. If you really want it, just have a goal, have perseverance, go for it. It’s okay to fail. Keep pushing, keep doing it millions of times, and then there will be one chance or one person who will believe in you and you will get it. And then you will pay it back for other people.

Want to learn about the ways that solutions architects are innovating for their customers? Check out Jamila’s blogs:

- Backend for Frontends Pattern
- Multi Region Deployment of AWS AppSync with DynamoDB Global Tables

Explore more content that celebrates the achievements of women in tech, such as:

How Women@Startups is building a community of women founders

Bonnie McClure — Wed, 22 Mar 2023 14:04:28 +0000

The concept of community—bringing people with common interests together (or people together at all)—took on a whole new meaning when the COVID-19 pandemic hit in 2020. For the founders of Women@Startups, Sree Singaraju and Priya Koratkar, the pandemic highlighted an even greater need to bring together one particular community: women startup founders.

Founded in 2020, Women@Startups aims to provide visibility and voice to the challenges that women founders and women in technology face. The group intends to create a sense of belonging and support for women in tech, working to address the inequities and challenges faced by women in the industry. Through employee and customer support functions, the organization provides coaching, leadership development support, and connections to organizations that support funding.

Crafting an equitable future for women in tech

Priya Koratkar, co-founder of Women@Startups

As working mothers from immigrant backgrounds, Sree and Priya had a taste of isolation prior to the national mandate. Before joining AWS, both women had years of experience in typically male-dominated industries, Sree in finance and Priya in the tech sector.

“At some point in our careers, we realized that we were, at many times, the only women in the room and the only people of color in the room. And that formed a lot of our perspectives of what an equitable future for women in tech should look like,” Priya says. “Work-life balance conversations or conversations around equity in the workplace were nonexistent.”

It was the founders’ roles at AWS, as Solutions Architect and Sales Leader, that gave them the tools to expand these ideas into what would become Women@Startups. As Priya says, “We both came into this with the understanding that if we are now in a leadership position in an organization like AWS, which is on the forefront of helping startups, we have a moral responsibility to make a way and path for others who are coming in.”

Engaging women founders in open conversation

Year one at Women@Startups was not without its challenges. But the organization chose to harness these obstacles, instead focusing on an initial goal of creating awareness. Priya says: “year one was a year of learning, a year of connection, and a year of building community.” In determining strategy, “we focused on a lot of these conversations about what are some of the challenges that we are facing as employees. What are the challenges that our customers who are women founders and underrepresented founders focused on?”

Sree Singaraju, co-founder of Women@Startups

With these challenges in mind, the first initiative from Women@Startups, re:Connect, was born. re:Connect connects a small group of women leaders in candid, open conversation. According to Sree, members discussed topics such as, “I’m a working mother, and I’m struggling right now with COVID because my kids are at home and I’m figuring out how to strike a work-life balance.” It was the answers that formed the basis of what the future programs at Women@Startups would look like. “We tried to make it as actionable and as practical as possible. So, we selected leaders who were not afraid to speak out the truth and be real with the candidates and with the community here,” Sree says.

re:Connect attendees

Creating a unique perspective for an employee resource group

Employee Resource Groups (ERGs) like Women@Startups are not new at Amazon. Amazon has 13 ERGs that unite employees with shared identities across the globe. Groups such as “Body Positive Peers” and “Indigenous at Amazon” support each other and promote diversity and inclusion within the workplace. While ERGs by nature tend to be internally focused, Women@Startups puts their focus on both the internal and external to cater to their customers and provide a unique perspective. “Ours is an ERG that has two customers that we serve: one is our internal employee group, and the other is our customers, the female startup founders. That’s the key differentiator between any other ERG and us,” Priya says.

Today, Women@Startups has grown into a global organization with over 170 team members and a substantial global presences. They have secured sponsors who serve as leaders across regions to help expand and bring visibility to the challenges that women founders face, including Howard Wright, VP of AWS for Startups; Paul Duffy, North American (NAMER) solutions architect leader; Sherry Karamdashti, NAMER sales director; Kellen O’Connor, Europe, Middle East, and Africa (EMEA) sales director; and Gaurav Arora, Asia-Pacific (APJ) sales director.

Women@Startups with Howard Wright, VP of AWS for Startups

With this growth, Women@Startups has stayed true to their core values, while expanding their reach. “Our goal is to make sure that we always are mindful about creating a pipeline of diverse candidates, so we can have a wider pie. Once we have those employees in, we ask: how do we continue to make sure they are successful in their environment and that they have the right tools to construct and grow?” Priya says.

Externally, on the customer side, Women@Startups looks for opportunities to make connections with organizations that are supporting women founders from a funding perspective. “We like to make those introductions so that we are not just looking at coaching opportunities for our women founders, but actual practical application of how they can be connected with the right folks who could give them funding.”

As leaders, Priya and Sree go by the “Amazonian” leadership principles: think big and invent and simplify. These tenets form the basis for how the team thinks about future growth. “We really encourage every member to come with a think-big strategy and a long-term strategy of the work that they’re doing,” Priya says.

Focusing on the future

The pair is certainly not done; with women founders, funding bias remains prevalent. In 2022, women-founded businesses only raised 1.9% of all venture capital funds, a drop from 2021. As Sree says, this forms the focus of the future for the organization: “Last year, we learned some of the challenges of women in tech. This year, we really want to dive deep and see what we can do in the tech funding space.”

Priya agrees: “We want AWS for Startups to be a destination for all women to grow and build, whether that is internal, as employees, or any woman-led startup. Any startup should think of AWS as their partner and as the ecosystem that they have for them to be really successful. AWS for Startups has the understanding and the empathy to support them in the ways they need.”

Envisioning an equitable future together

Explore more content that celebrates the achievements of women in tech, such as:

Greylock’s Jerry Chen on the evolving role of the startup CFO

Kate Radinovic — Tue, 21 Mar 2023 14:04:00 +0000

Welcome to our new thought leadership spotlight, “The Evolving Role of the Startup CFO.” This series features perspectives from prominent players in the startup ecosystem. These blog posts tackle critical questions, including: What does the role of today’s startup chief financial officer (CFO) entail and how will it evolve over the lifecycle of a startup? How can we most effectively support CFOs as the cloud increases its dominance within the organization and balance sheet? And can the CFO better navigate—and ultimately enable—the relationship between technical leaders, CTOs, and engineering teams?

Leading our first spotlight is Jerry Chen, a veteran partner at Greylock Partners, a globally recognized venture capital firm.

The modern startup CFO plays the role of team quarterback

“The CFO really ties it all together,” says Jerry, stressing that the CFO’s responsibilities go well beyond simply saying “no,” which is how the role is often stereotyped. Instead, a dynamic CFO should serve as a partner to the chief executive officer (CEO), founders, and other company leaders—including those in charge of sales, marketing, product/growth analytics, research and development (R&D), and engineering. “The CFO, if he or she is doing her job right, is playing quarterback between all those departments.”

A strategic CFO should be familiar with—and understand which key levers drive—their company’s business model. Above all, it’s important for the CFO to understand which metrics matter most to their own company.

“A great strategic CFO can boil down the four or five metrics that matter for the company as a whole, and then work with each department to understand, from the engineering side, the drivers of cost from a people-cost perspective,” says Jerry.

When a CFO understands the key business drivers, they can better work with executive and people leaders at the company to ensure that they’re making smart hiring and salary decisions. “You build your board for good times and bad times,” says Jerry. Similarly, “Make sure your CFO is going to be a partner in crime to help the startup weather whatever bumps lie in the road ahead.”

Cash is king—and metrics matter

When it comes to non-negotiable metrics, Jerry says CFOs must prioritize the obvious ones. “What are the three reasons startups go out of business?” asks Jerry, who then paraphrases former professor Bill Sahlman from Harvard Business School. “One, they don’t raise enough cash. Two, they don’t price their solutions correctly and ultimately are not collecting enough revenue from customers. Three, they spend too much cash. Startups die because they run out of cash. So, it’s really important to focus on the cash burn, cash flow, cash burn rate.”

In addition, CFOs must pay attention to the company’s gross margins and watch the trend line of whichever metric is most important to their business. “A point in time doesn’t matter to your CFO or to the CEO,” says Jerry.

“It’s the trend line that matters. Are you going up and down? And more important than the trend line is the second derivative. Is burn increasing or decreasing? Is revenue growing faster than before? What’s the long-term gross margin?”

Introducing a strategic CFO to your business at the right time

It’s important to get timing right when hiring a CFO. “The CFO title comes later in life, when a company is pretty far along,” says Jerry. “CFO or C titles are typically reserved for companies at scale.” In the early days, when a startup is only dealing with expenses, they will most likely be able to outsource their payroll to a third party. When the company starts generating revenue, they may need to hire a director of finance in order to manage both the sales and revenue.

“As you’re scaling up, whatever the top line is, you’re going to need a vice president (VP) of finance who can grow into a CFO at the right size of scale,” says Jerry. “Then you have a CFO that needs to play central clearinghouse for the CEO because the growth team, the marketing team, the sales team, and the product team will all have their own metrics…And then, the bonus level is if the CFO can be a strategic partner to the CEO and the founders and the board about what’s right or wrong. That’s next-level.”

The CFO has the opportunity to instill a culture of collaboration, where everyone, especially the startup executives, are working together to reach a common goal. “I think in any healthy community culture, having someone who’s keeping an eye on the bank account is useful,” says Jerry. Sometimes the CFO may need to play devil’s advocate, reminding everyone else that while they’d love to build everything with infinite time and resources, they can’t because the company needs to prioritize gross margins.

“The CFO’s not going to say yay or nay, but should give the resources and the framework for the CEO and the founder to make the decision.”

If your startup is looking to hire a great CFO, past performance is not the only important attribute. Jerry says that when ranking key attributes, agility and adaptability may prove to be the most important for a startup CFO.

“At a startup, your job’s constantly changing. The CFO’s job in 2023 is going to be different in 2024 and will be different in 2025,” says Jerry. “Hopefully you’re going to hire somebody that has the mental agility to grow through those roles.”

Women founders Q&A – Learn how they’re impacting their communities, industries, and beyond

Bonnie McClure — Wed, 15 Mar 2023 18:04:43 +0000

Today, we’re talking to six women founders and leaders about how they’re making impacts in their communities, industries, and beyond.

Ritu Chakrawarty, founder of Graaphene, a solution for last-minute backup childcare. Graaphene is a mobile app that connects parents with trusted, vetted caregivers on demand.
Caitlin Colgrove, co-founder and CEO of Hex, a platform for collaborative data science and analytics.
Veronica Falzone, co-founder and CEO of Thumbo, live sports audience engagement software that helps teams improve their fan experience, while collecting actionable data.
Leanne Linsky, founder and CEO of Plauzzable, an online comedy platform that offers live stand-up comedy for their fans.
Anna London, co-founder and CEO of Chrysallis AI, a scenario-based learning platform with gamification powered by AI.
Barr Moses, co-founder and CEO of Monte Carlo, a digital data reliability platform designed to monitor and offer alerts for missing or inaccurate data.

First up, Ritu and Caitlin tell us how the startup environment has changed in the few year since they both got started.

When you compare the startup environment for women founders today to what it was like when you were first getting started, what has changed?

Ritu, Graaphene – It’s remarkable how much the entrepreneurial landscape has changed in such a short time. The environment today is vastly more supportive than it was even in late 2020. Thanks to the efforts of various organizations dedicated to promoting and nurturing female startup talent, there are now many resources available, including incubators, accelerators, female-focused investment funds, and VCs, as well as advocacy and mentoring organizations.

Ritu Chakrawarty, founder of Graaphene

However, despite these developments, inherent biases are still holding us back. A recent report by HBR highlights the kind of obstacles women-led startups may face. Shockingly, women-led firms still receive less than 3% of all VC investments.

To counter this, there has been a push to get more women involved in venture financing. Studies show that female investors are more likely to invest in female founders. But, as shown in this research, having support from female investors could actually make it tougher for female founders to raise more money down the line. Attribution bias is a big deal. When people see a female founder getting funding from a male investor, they think it’s because she’s competent and her startup is strong. But, if that same founder only has female investors, people are more likely to assume her success is due to her gender, not her competence.

Although we have made a leap and many of the recent developments are extremely positive and represent a significant step forward for female entrepreneurs, we still need to work together to be aware of these biases so we can create a more fair and equal entrepreneurial landscape for everyone.

Caitlin Colgrove, co-founder and CEO of Hex

Caitlin, Hex – My first few years out of college I really remember as peak “brogrammer” culture – it was one of a number of toxic traits that rapidly scaling startups often developed back then, many times unintentionally. It used to feel like you had to fit a certain mold to work at a startup (much less start one), and it was one that I personally never fit.

Over the years though, I’ve been really excited to see how many startups are thinking about building culture from day one, and the result has been that today you can find early stage companies that are welcoming of all sorts of personalities and backgrounds. Without some of those examples to follow, I honestly don’t know if I would have chosen to be a founder. We’re definitely far from perfect as a company or as an industry, but this is one trend I really hope continues.

Next, tell us about some of the obstacles you’ve faced as a founder. What did you learn about yourself as you confronted them?

Ritu, Graaphene – One of the most significant hurdles I encountered while bootstrapping the entire process at Graaphene was the lack of venture capital financing, which meant we had to run on a tight budget.

The one thing that I learnt about myself embracing failure with grace, having excelled throughout my academic and corporate career, learning to hear “no” wasn’t easy. However, in the startup world, rejection is inevitable, and I’ve learned to use each “no” as an opportunity to learn, pivot, and refine our approach. Staying laser-focused on our “why” and remaining committed in the face of setbacks and difficult circumstances has been the antidote to overcoming these challenges.

Veronica Falzone, co-founder and CEO

Veronica, Thumbo – One obstacle I have faced is finding quality talent to join our team. A strategy that has helped is getting to know the person beyond their qualifications, through asking questions that help me understand their passion for our company’s vision. A passionate team member is one who will have sustainable focus and drive, not just at the beginning, but all the time, because they care about our mission.

Another challenge I have faced is learning to be patient. For founders, there is not the same instant gratification that comes in some other roles. What I’ve learned about myself is that when I focus on self-created daily benchmarks, I can stay motivated and on goal every single day.

Lastly, as a woman CEO I have had to deal with overt sexism. Several people I’ve encountered assume my male co-founder is the CEO before even speaking with us. I’ve learned that I am at my best when I don’t let that deter me, instead using as fuel to ensure my voice is heard. I’ve also learned to prioritize speaking with diverse and like-minded mentors and potential investors.

Leanne Linsky, founder and CEO of Plauzzable

Leanne, Plauzzable – Feedback. Everyone wants to give feedback to a founder. Filtering all the feedback and advice can be not only time-consuming, but confusing. I’ve learned that customer feedback always gives me the most helpful insight into my desired results. I ask open-ended questions and let the customer talk. Anytime we question features or user experience (UX), I can count on my customer to give me the right answer.

Anna, Chrysallis.ai – One of the biggest obstacles I’ve faced is limiting beliefs. Mindset is everything. Being a woman is only a limiting factor if you allow it to be. Where there is not an opportunity for yourself, create one. Don’t wait to get asked to the table. Pull up your own seat and sit there and let your voice be heard.

Anna London, co-founder and CEO of Chrysallis.AI

At one point, I was told to add a male cofounder to the team to attract investors, as only 2% of funding goes to female-led startups. I was also told to put a man in charge, because a male-run company is perceived as having lower risk. Instead, I chose a female cofounder. Not because she is a female; rather, because she has the knowledge, skills, and abilities to get the job done.

Together, she and I developed and launched a beta and have been able to drive hundreds of users to the platform. We’re looking forward to setting the example for how female-led companies can help change the mindset of investors and venture capitalists and pave the way for more women-led tech companies to be the catalysts for innovation and transformation.

Also, what makes you excited about or gives you hope for the future of women-owned and led startups?

Ritu, Graaphene – We have the momentum to create a more equitable world that our upcoming generations deserve, and it all starts with acknowledging and celebrating the contributions of female founders. Founders are a special breed, requiring courage and tenacity to bring a new idea to fruition.

But when it comes to exceptionalism, women founders stand out. They not only take on the same multi-faceted roles as male founders, but they also face unique challenges. Limited investment dollars mean that they must bootstrap their startups while juggling household and caregiving responsibilities, often starting later in life. This extra power to overcome adversity, combined with fresh perspectives and innovative strategies, makes female founders invaluable and exceptional.

There is a growing awareness of what value female owned businesses bring, and that’s driving a shift towards supporting diverse talent. Incfile’s recent report reveals that for the first time, women entrepreneurs are growing at a rate that outpaces their male counterparts by over 20%, with a 76% growth rate among women over 65. This inspiring data gives me hope that we will soon see a more gender-equitable world.

Barr Moses, co-founder and CEO of Monte Carlo

Barr, Monte Carlo Data – In 2023, more than 10% of Fortune 500 companies are led by women – an all-time high. It’s a really exciting time to be an entrepreneur, and I’ve been so inspired by the founders and operators I’ve had the pleasure to meet in the AWS community. There are a lot of problems out there to be solved, and I have full confidence that these individuals will be at the forefront of the next generation of industry-leading companies. At Monte Carlo, for instance, I have been privileged to work with some of the best women in their field, and we’ve only just gotten started scaling and leading the data observability category.

Finally, what advice do you have for young founders who want to start their own company?

Barr, Monte Carlo Data – As a startup founder, speed is your biggest advantage and focus is your biggest challenge. There are always a million things you could be doing, and just a handful of areas where your time and attention matters most – lean into those areas, move quickly, and scale them as far as you can before moving onto the next area. Ask yourself what will it take to get something done today instead of tomorrow, for instance: “What will it take to ship this new feature this week, instead of five weeks from now?” At Monte Carlo, our two operating principles are “speed and focus” for this very reason.

Caitlin, Hex – This journey is not for the faint of heart. You’ll face countless rejections and setbacks, and you’ll have to summon the strength to say “no” many times before you finally hear a “yes”. It will test you physically, mentally, emotionally, and spiritually. But if you’re truly obsessed with solving a problem, if it’s something that you feel in your bones is worth fighting for, then you’ll be able to endure the challenges and stay the course.

Leanne, Plauzzable – Never be “above the broom.” It’s the opposite of the “It’s not my job” mentality. For example, if the floor needs a quick sweep, I pick up the broom and sweep it. If I answer the phone and a customer asks a question, I will help them.

Don’t make your customer step into a pile of construction debris because using a broom is not written in your job description. Don’t make your customer go through the process of being transferred to 2 or 3 other people because you don’t work in customer service. Take the call and listen to the customer’s question. If you don’t have an answer, go find one and come back to the customer with it. A good leader shouldn’t ask of others what they aren’t willing to do themselves.

Ritu, Graaphene – Focus on PCP – Problem, Customer, Perseverance. And when it comes to perseverance, Steve Jobs once said: “You have to be burning with an idea, or a problem, or a wrong that you want to right. If you’re not passionate enough from the start, you’ll never stick it out.”

For technical founders: do a stint as an engineering manager. Learn hiring, performance management, etc. Engineering management is really hard, and it’s even harder if you have to learn how to be a manager at the same time as learning how to be a founder. I’m really grateful for my management background because once we started to gain traction it really helped me accelerate building a really strong engineering team.

Explore more content that celebrates the achievements of women in tech, such as:

- AWS Startups Twitter stories that highlight amazing women in tech
- Celebrating Women in Tech with Panzura CEO Jill Stelfox
- Demo Day for AWS Impact Accelerator Women Founders Cohort
- Meet three women who make startup journeys smoother
- How Women@Startups is building a community of women founders

Demo Day for AWS Impact Accelerator – Pitching to 40+ investors

Hannah Perry — Wed, 08 Mar 2023 17:44:01 +0000

Have you ever wondered what it’s like to stand on stage at the Demo Day for the AWS Impact Accelerator? With over 40 investors in the audience, you might start to sweat, but the AWS Impact Accelerator Women Founder cohort made it look easy! Participating in Demo Day is the challenging (if nerve-wracking) culmination of 8 weeks of building, technical investment, mentorship, and hard work.

At AWS, we know that talent is everywhere, but opportunity is not. The AWS Impact Accelerator is changing that narrative. Combining the power of Amazon and AWS technology, we’re helping women founders accelerate their cloud-based business for serious scaling.

In September, 25 startups were selected to participate in the in the AWS Impact Accelerator for Women Founders cohort. Each of these startups received up to $225,000 in cash and AWS Activate credits, an extensive and individually curated training curriculum, mentoring and technical guidance, introductions to Amazon leaders and teams, networking opportunities with potential investors, and ongoing advisory support.

After a week-long kickoff in Seattle, founders bonded through their virtual community, participating in weekly workshops and one-on-one technical and business trainings in partnership with a personally curated mentor. In December, the founders came together again at the AWS Startup Loft in San Francisco to prepare their final pitches to a diverse group of 41 investors from 31 firms.

Check out the video to get a look at the fun, and learn about all the AWS Impact Accelerator opportunities currently available to startups through AWS.

Explore more content that celebrates the achievements of women in tech, such as:

How building on AWS made a big pivot possible for syniotec

Bonnie McClure — Tue, 07 Mar 2023 17:14:05 +0000

A few months after syniotec launched its first product, the team realized they had a problem. The Germany-based startup was ready to revolutionize the construction industry by becoming the Airbnb of construction rentals. They wanted to help machinery owners around the world optimize efficiency by renting out their equipment that wasn’t currently in use.

As it turned out, their product was ready—but their customers were not. Co-founder and Chief Operating Officer Rezi Chikviladze fielded early feedback from construction clients who were excited about the possibilities of the service, but saw it as too futuristic for an industry not known for its digitization.

“We don’t know where our machines are, how our machines are planned, whether the construction site managers need those machines,” was the refrain Rezi kept hearing from potential customers.

From there, syniotec went on to find out that not only did potential customers not have tracking devices on their assets, but many lacked the digital infrastructure to implement a software solution in the relevant working areas. It wasn’t uncommon for Rezi to chat with construction companies who used old systems like Excel-based lists to dispatch equipment to job sites.

Customers may have been eager about the cost saving potential of renting out their unused assets, but they simply did not have the systems or processes to accommodate syniotec’s idea. The company came to a new conclusion: Before becoming the Airbnb of machinery rentals, they had to build the basis that would help construction companies into a new era of digitization.

Building a new foundation

After that, syniotec began evaluating the most urgent problems construction companies needed to solve. They heard from potential customers who were managing as many as 200,000 pieces of equipment in more than 25 countries.

These giant companies had no centralized system to manage their fleets, keep up with complex scheduling issues, and track maintenance and compliance checks. It wasn’t uncommon for companies to completely lose track of entire pieces of equipment, or be unequipped to optimize their fleet’s usage.

“The difference between planning and real usage on the construction site is huge,” said Rezi. “And at other times, machines are standing there on the site not utilized, and this is of course a financial loss for the company.”

Those losses can make a huge dent in a company’s reputation—and can add up quickly. So, syniotec customers were eager for solutions that optimized costs by even a small percentage, since that still had the potential to make a giant impact on bottom lines.

Introducing SAM

syniotec packaged their solution as a Smart Asset Manager they dubbed SAM. SAM is a web and mobile app designed to manage all the moving parts of fleet management in one, centralized location. Construction companies can use SAM to dispatch construction equipment to proper locations, track where any piece of equipment is and monitor its usage or inactivity, more accurately plan equipment scheduling and manage machinery transport.

A drastic reduction in dispatch call times has been one feature of SAM that excites customers, says Rezi. Previously, when calls came in to request for a certain machine at a construction site, it could take the dispatchers a lot of time to sort through paperwork or slow systems in order to correctly assign out a piece of equipment to a site. But now, dispatchers utilizing SAM have been able to decrease the average call from 30 minutes to just three.

Regulatory and technical checks are also more efficient with SAM. Rather than tedious equipment checks and mountains of paperwork, customers can simply scan their machine with their phone and upload and store the relevant data to their AWS profile.

SAM can even work as a people manager. The unified system makes it easier for companies to calculate hours worked on site, acting as a type of enterprise resource planning (ERP) platform that helps to cut costs and reduce the time it previously took to manage payment calculations.

Additionally, syniotec is leveraging Internet of Things (IoT) capabilities to help construction companies have constant eyes on their equipment, no matter where it is located worldwide. By using the IoT via a small electronic device, companies can easily collect data on every operating aspect of their fleet.

Details like current voltage and the hours worked at specific sites can give companies the tools they need to stay on top of routine maintenance, be prepared for compliance checks, and ultimately save tons of money making sure each piece of equipment is operating at its top capacity. Thanks to the newly acquired IoT data, one customer was even able to finally track down equipment that had been stolen—right down to the garage where it sat—helping the police to find stolen goods worth more than €300,000.

Migration from monolith

syniotec’s pivot wasn’t only a transformation on the business side—it also meant a complete overhaul of the technical infrastructure that powered their product. The company’s original rental facilitator idea could have been served by monolith architecture. But as they transformed their business and began offering a far wider variety of services to their customers, syniotec recognized the need to switch to a microservices architecture that could offer greater agility and scalability.

The pivot was only possible with AWS, Rezi says. Building elsewhere “would have cost us a huge amount of resources to make such a huge change fundamentally.” But since they had used AWS for their monolith, already had the support of the AWS team, and could choose from such a large portfolio of offerings, syniotec could transition with ease.

Newly situated on Amazon Elastic Kubernetes Service (Amazon EKS), syniotec was far better positioned for reliable scaling. With the Kubernetes auto scaling group and Amazon CloudWatch, the team’s developers could better observe their microservices’ behavior, ensuring a more transparent and efficient process.

In doing so, they could also be more on top of any system issues, and hoped to minimize their response and resolution times. They considered configuring elastic search on their own, but found that using Amazon OpenSearch Service allowed them to save tons of time and resources. Plus, using OpenSearch Service gave them the peace of mind that their critical operational data was backed by the AWS commitment to security.

That commitment to security is also on display with syniotec’s backend services. By using a Amazon Virtual Private Cloud (Amazon VPC) link and an Elastic Load Balancing (ELB) Network Load Balancer, AWS connects the company’s backend services in a private network only accessible via Amazon API Gateway. The team can create private integrations and custom domain names to make obtaining and renewing certificates simpler and more affordable.

With the increased ability to scale, the syniotec team is now managing over 50,000 assets. This means managing the live operative data coming in from its IoT telematics solutions connected to construction equipment worldwide, plus data collected directly from equipment managers. Additionally, the company receives more than 2.5 million messages in a single day.

Keeping up with that volume of data and requests at speeds necessary to accommodate the construction industry was not an easy task. But syniotec has found the answer with Amazon Simple Queue Service (Amazon SQS). The team chose it for its high reliability, safety, and performance speed, and has found it a necessary tool, especially when they need to handle increased demands in a short period of time.

Along with making their customers happy and their jobs easier, transitioning away from monolith architecture had a huge impact on syniotec’s bottom line. Thanks to increased productivity and efficiency, the switch to Amazon EKS meant lowering their provisioning time and slashing costs by half.

Paving new paths worldwide

Now that syniotec has overhauled and scaled their infrastructure and can offer customers the solutions they need in the current construction climate, they are looking ahead to new ways to drive digitization in the industry.

But first, they’re focusing on expanding their suite of solutions that meet current industry needs. The team is looking forward to exponential growth both in their current operating countries of Germany, Austria, and Latvia, as well as beyond those borders.

The company is aware of the great need they will be able to fill as the construction landscape evolves. More companies and regulatory bodies are now pushing for sustainable and innovative building to accommodate a warming planet and a growing population, including moving to electric equipment. But without more digital tools, many construction companies will be unequipped to provide the level of transparency and efficiency that evolution would require.

syniotec, then, is ready to become the indispensable platform its customers will rely on to make more data-driven decisions on their fleet and resource management. By using AWS to leverage the power of IoT and advanced technology, syniotec is helping construction companies worldwide synchronize their business with the possibilities of the future.

Applications are open for the AWS Impact Accelerator Latino Founders cohort

Keely O'Neill — Mon, 06 Mar 2023 14:51:31 +0000

AWS is launching the next Impact Accelerator cohort, giving pre-seed Latino founders the support they need to accelerate their businesses.

It’s almost been a full year since the AWS Impact Accelerator launch—a $30 million fund that provides Black, Latino, women, and LGBTQIA+ founders with equitable access to funds, training, mentorship, tools, and resources.

Today, AWS launched applications for the next cohort—Latino Founders. While Latinos represent a key component in the makeup of the US population, they are still among the communities that receive less funding and support for venture capital. Even at its peak in 2021, US venture invested only 2.5% ($8.5 billion) in Latino-founded companies, according to Crunchbase. Through Q3 2022, funding to US-based Latino-led companies dropped to $2.7 billion, which is only 1.5% of US venture dollars. This pullback in funding to Latino founders in the US took a sharper downward turn than the broader slowdown in venture capital in the US market, which was down more than 50% year over year for funding in the same timeframe.

On the other side of the table, Latino investors make up only 2% of the venture capital industry relative to 19% of the US population, according to LatinxVC’s State of the Latino/a VCs annual report. By partnering with Latino-led venture capital firms like LatinxVC and VC Familia, the AWS Impact Accelerator is providing programming and mentorship to help Latino founders overcome bias and lack of representation prevalent in the venture community.

Lolita Taub is a Latina General Partner at Ganas Ventures, investing in community-driven companies and changing the face of the startup-venture capital tech world. When asked why this mission is so important to her, she reiterated that we need more resources and programs to support Latino-led startup founders and fund managers, adding,

“The Latino market is equivalent to the 7th largest economy in the world and represents over $2.3 trillion in new opportunity for our community and our wallet!”

Tailored benefits

Over the course of eight weeks, selected startups accelerate growth by developing their ventures alongside AWS technology experts, investors, and partners. To further their growth, startups receive a $125,000 unrestricted cash grant and $100,000 in AWS Activate credits—all at zero cost and for zero equity. Startups also gain access to the alumni community of 50 previous AWS Impact Accelerator participant startups and over 200 founders, CEOs, CTOs, and mentors from the Black Founder and Women Founder cohorts.

Members will learn how to navigate the fundraising process with sessions such as Overcoming Bias in Fundraising and Effective Storytelling on the Path to Raising My Seed Round. Through these sessions and more, they’ll hear firsthand from those that have successfully raised funding rounds despite the lack of representation or investors that may not identify with their market. Lolita Taub will also lead a session about building investor relationships and pipeline, and share her perspective as one of the investors looking to change the current funding disparities.

How to apply

New for the AWS Impact Accelerator Latino Founders cohort, applications are open to founders located in the US and Latin American countries, as long as their startup is incorporated in the US If you’re interested in applying, get started early. Applications are open March 6 – March 17, 2023, with limited space available. Submit your application here, and don’t forget to review the FAQ and terms and conditions for the full list of qualifying criteria.

For those new to the accelerator application process, we’re hosting two in-person informational sessions at the AWS Startup Lofts during launch week. Join to hear from the AWS Impact Accelerator team, understand timelines and expectations, and meet with alumni members.

San Francisco Loft | Thursday, March 9 at 3:00 PM PT
New York Loft | Thursday, March 9 at 3:00 PM ET

Apply today

The AWS Impact Accelerator gives high-potential, pre-seed startups the tools and knowledge to reach key milestones, such as raising funds or acceptance into a seed-stage accelerator program, while creating powerful solutions in the cloud.

Applications are now open for Latino Founders seeking to accelerate their startup’s growth from 0 to 60 in eight weeks. Apply online today.

Overcoming criticism and imposter syndrome as a startup founder

Hannah Wall — Fri, 03 Mar 2023 19:00:16 +0000

When you think of startup culture, stories of meteoric successes leap to mind. But for every high-flying unicorn, you may be reminded of the risks and failures also associated with startup culture. Founders know that with success comes detractors, but how can you manage your own self-confidence and sort criticism from feedback? Identifying, and more importantly, managing your approach to criticism is key in maintaining positive mindsets and positive culture within a startup.

According to new research commissioned by AWS, “tall poppy syndrome” — a cultural attitude that disapproves of success and seeks to cut down those who stand out — is widespread within the startup ecosystem, with 80% of surveyed startup leaders saying they have experienced it personally. This is particularly true among the new generation of founders; leaders of younger startups are more likely to have experienced tall poppy syndrome than their counterparts, the research shows.

Social media is attributed as being a leading source of criticism driving “tall poppy” feelings, but it can also come from friends, loved ones, and other founders.

“It’s the mother-in-law or the friend or the Instagram post that is constantly questioning whether you have done the right thing,” says Barb Hyman, founder and CEO of the artificial intelligence (AI)-based human resources (HR) and hiring startup Sapia.ai.

Such criticism can stifle a startup’s growth. Nearly all startup leaders who have experienced tall poppy syndrome believe it hinders growth potential, with 46% saying it has caused them to be more risk-averse, 45% saying it has hindered career development, and 43% saying it has caused their mental and emotional well being to deteriorate.

The result: founders are reluctant to declare themselves to be successful; they would rather keep the focus on the success of their teams.

Barb also says that founders should stay true to their vision in the face of such criticism and remind themselves that what they are doing is important.

“I just love proving people wrong, and especially those who don’t believe in what I can do,” says Barb. “Every day I feel like I am learning, and helping, and ideating. What we are doing is disrupting the whole way we think about people, so it is fundamentally creative, and we are solving so many things because of that.”

And sometimes the loudest detractor in the room? You. Imposter syndrome, or the feeling that accomplishments are the product of luck over ability, is common to most founders — three-quarters of startup leaders say they experience feelings of imposter syndrome, with one in eight feeling it daily. But it’s something that can be managed, experts say.

More than one in four startup leaders we talked to say they manage imposter syndrome in positive ways by focusing on building resilience, celebrating success or cultivating self-compassion.

“I typically manage imposter syndrome feelings by speaking openly with my team,” says Dr. Ben Hurst, founder and CEO of the patient engagement platform HotDoc. “I believe it’s important not to set unrealistic expectations regarding my own capabilities. I am just as fallible as everyone else, trying my best and doing a lot of this stuff for the first time and hoping to improve myself by learning from my team.”

Not addressing feelings of imposter syndrome can have more detrimental effects. 21% of startup founders admitted to working until burnout, and 19% say they would like additional help in managing these feelings.

Hurst says that founders shouldn’t be afraid to be open and honest with their teams about what they are able to achieve, however. “Our core values are always to be empathetic, take ownership, and speak up, and those were architected based on what makes for a successful doctor/patient relationship,” Hurst says. “The most important thing is authenticity and ‘show, don’t tell.’ If I preach these values and am not seen to uphold them, then suddenly it all comes crashing down.”

The takeaway? The journey to founding your startup may feel lonely, but you aren’t alone. Sharing your feelings and celebrating your successes, no matter how small can help you cultivate the self-compassion needed to weather criticism and doubt. AWS knows the right team and support network can make all the difference, which is why we have helped more startups build, scale, and succeed than any other cloud provider.

Check out the rest of the research and insights in the report, “How startups create a culture of possibilities,” to learn from established founders about what it takes to create a culture for success and the key factors and considerations that have made the difference as they’ve grown.

Celebrating Women in Tech with Panzura CEO Jill Stelfox

Jill Stelfox — Wed, 01 Mar 2023 14:56:48 +0000

Jill Stelfox, CEO and executive chair of Panzura

Meet Jill Stelfox. Jill is a serial founder and entrepreneur, holds multiple patents, and is the chief executive officer (CEO) and executive chair of Panzura. Panzura’s award-winning CloudFS global file system gives the ability to access files from anywhere with visibility, security, and control.

Alongside leading Panzura to successful new heights, Jill actively celebrates and elevates women in tech. She co-founded Women in Sports Tech and is a sitting member of their board.

“One thing I’ve seen is that women want to see it in order to be it,” says Jill. “They want to know that it’s possible. Sharing stories about other women in tech can show them that it’s totally possible. Look, if a grandma of two can be a CEO, you can too.”

Can you please introduce yourself, as a person and as a professional?

I’m Jill Stelfox. I am a mother of two and a grandmother of two: a one-year-old and a three-year-old. I am happily married for 33 years. Great family life. I am also the CEO of Panzura.

Why is Women’s History Month important to you?

Women’s History Month is important to me because while women have come far in terms of equality—especially in the workplace—there is still a ways to go.

As a woman in tech, I have raised a couple hundred million dollars’ worth of venture capital and private equity, and I have returned billions to investors. Yet, I am often the only female CEO in an industry and certainly in a room.

Panzura was refounded under your leadership, resulting in many successes on both the technical and cultural fronts. What are the challenges and the opportunities of “refounding” for startups?

Refounding a startup is challenging because the premise of it is you take the core of the goodness that’s there and you change it. We refounded the company three years ago and since then we’ve grown 485%. Panzura has tripled the number of products that we have in the market. We’ve added hundreds of customers.

Helping the team that was here understand that the transformation is going to lead to greatness takes a lot of handholding and care in the very beginning. All kinds of great things have happened through really seeing the possibility of hybrid cloud technology and what it can do in terms of understanding what’s going on with your data for customers.

It’s been hard work and tons of fun.

Can you tell us about a Panzura success story for which your experience as a woman CEO and veteran strategist was critical?

I’ll never forget the very first day that we took over Panzura. It was May 7th, 2020, and those were the early days of Covid. Imagine getting on a web call during that time and announcing, “The company’s been purchased and I’m your new CEO.” People were filled with the fear and uncertainty of everything. Some people got teary-eyed about it.

It instantly changed me forever.

It’s a wonderful time for women to run companies because we can usually see when people are hurting. Men can too, but we have more freedom as women to be able to do something about it. So it was easy for me to be empathetic and to be “the caring person.” That is my nature.

I’d had all these fancy notes about what I was going to talk about and how I was going to change the technology … how we were going to do all these great things and go after new customers and new markets. Honestly, none of it mattered on that day. On that day, what mattered was being human—talking about how scary this was, and what would “being okay” look like in terms of the company.

When I hung up the call, everybody seemed happier. They were in a calm place. I hung up that call and thought, “I need to do this really well for these people. They deserve comfort and they deserve kindness.” My mission became not about the tech, but about the team.

How has AWS supported your goals for Panzura?

Our AWS relationship started through some folks approaching me—and they were women—about what an AWS partnership could mean to the company.

We had raised some new money and the new investors that we brought on were also women.

The night that we signed the agreement with AWS, which was for us a company-changing agreement, we celebrated with a dinner. At that dinner, we realized that the entire transaction was done by women. It wasn’t because we were trying to do something gender-based; it’s because we were all in positions where we could affect change in this way.

I’ll never forget in my whole career that the first time I did a multimillion-dollar transaction, it was all with women and it was AWS and it’s outstanding.

Celebrating the agreement between Panzura and AWS at dinner

To answer your question about how does AWS support our goals—when we refounded Panzura, we switched from selling to mid-size companies to selling to large-scale enterprises. There are some of the largest financial institutions, healthcare institutions, and manufacturing companies in the world. We work with them to take their data workloads that are on-premises servers to the cloud in a really effective way.

When we do that, we can reach out to our AWS team and get instant access to the AWS team member of that specific customer account. It’s amazing. It’s literally one phone call and we have the right people in the right meeting with the right attitude.

We are so aligned philosophically with AWS on the mission of bringing these important workloads to the cloud and really helping these companies be more effective in what they’re doing to make their companies successful.

The second part about working with AWS that’s great is our technology runs through AWS. As an AWS Partner, we have a technology stack that’s available in the AWS marketplace.

So it’s a two-part relationship with AWS: technology and go-to market. On the technology side, we have met some of the brightest thinkers. They’re so invested in our success. You get in these rooms where you’re talking about solving a technology problem. It’s an AWS person and Panzura people, and the exchange is so free and open and productive and useful. It’s a great relationship in that way.

Can you share some advice for women founders?

Yes. I have always had a coach. When you play sports, it’s just so natural that you would have a coach. In business, you need a coach, too. There always has to be somebody that’s on your side to cheer you on.

These are not expensive resources, and they can be found at any level. It’s just a matter of reaching out. Sometimes we as women don’t ask for that kind of help, but it is there to be given.

How do you envision the future of women in technology?

The future of women in tech, in my opinion, is that we represent at least 50% of those in tech.

We’re certainly capable of doing it and we’re capable of doing it along with managing everything else in our lives: our families and our homes, our pets, and our spouses. We can do all the things and still be capable of being equal at the table.

We have made a lot of strides. When I first got into tech 25 years ago, there were no women around at all. Now I see women at all levels. I don’t see them very often at the C-Suite. But definitely, if you look at our board, our board is 50% women, our company is 30% women.

We have come a long, long way. And I think we have more to go, but we’ve come a long way. We need to raise each other up as women, we need to hire each other. We need to boost each other, we need to promote each other. We need to do all of those things.

What’s next for you, for Panzura, and for Panzura with AWS?

On a personal note, hopefully I have a third grandchild soon.

On a work note, I could not be more honored to be the CEO of Panzura. I want to keep being the best one that I can be for this wonderful team that we’ve created.

We are determined to change, along with AWS, an industry that hasn’t changed in 20 years. There is a more efficient and safe way to store, locate, and manage your data. We want to be part of that revolution, along with AWS. We have made our strategic commitment as a company to be all-in on AWS. We are super excited about the possibilities of really changing the world.

Explore more content that celebrates the achievements of women in tech, such as:

Evolutionary architectures series, part 2

Justin Plock — Tue, 28 Feb 2023 13:30:26 +0000

The company’s beta launch is a success.

“I think we may be onto something.”

The first blog describes how Example Startup reached their first major milestone by delivering a minimum viable product (MVP). In part 2, we will see how Example Startup continues evolving their solutions to meet an increase in requirements and growth.

Building on the success of the beta launch

Things are starting to look up for Example Startup. The launch of their first MVP was a huge success for two reasons:

The number of people that signed up for the beta cohort of fantasy investors grew exponentially after word about the product got out on social media.
The startup got their first sponsors to chip in with some nice rewards for the winners of the beta cohort.

It is clear that the founders are onto something. Now, the startup needs some help before the next cohort begins and the company gets its first paying customers. It’s time to start hiring. Example Startup needs engineers that can take over the platform development while the founders pivot to leadership roles and start taking care of everything needed to get their startup to the next stage.

The great news from Amazon Web Services (AWS) couldn’t come at a better time. Example Startup is accepted into the AWS Activate program which means they can now access free credits to cover their growing cloud expenses. This give them some much needed runway. While the credits are much appreciated, the AWS Activate program also includes a number of other perks like a Premium Support Plan, as well as a relationship with an AWS account team that put technical and business expertise directly at their disposal.

With a couple of engineers joining the team, it is time to evaluate the solution that got them through the MVP and to begin planning for the next release. The technical founder starts the handoff to the engineers, which spurs many discussions around what went well and what needs more work. After documenting all the existing needs, gaps and questions, the team feels lost. There are so many options, so many decisions to make, and so little time. The technical founder decides it is time to talk to AWS again for some guidance.

Enabling growth with more AWS services and features

One of the first things on Example Startup’s list is business reporting. During the beta period, the founders didn’t have much insight into metrics like user signups that would give them a better sense of how their beta release was going.

The AWS solutions architect suggests Amazon QuickSight – a cloud-native, serverless business intelligence (BI) service. QuickSight has the ability to seamlessly integrate with their current database but also other data source they might need such as raw data in Amazon S3 or even data from external 3rd party providers. Building their first dashboards is a breeze with the user-friendly web interface allows them to quickly iterate to build what they want to see. Features like scheduled email reports allow them to wake up every morning with all the important information already in their e-mail inboxes. QuickSight also boasts threshold alerts that inform the team whenever any new milestones in subscriptions are achieved. What initially seemed like a huge undertaking was resolved in a matter of days.

The next big-ticket item for the team is accepting payments. This is something no one in the team has experience with. Following a couple of informative sessions with the AWS team, the team has a well-defined set of requirements that they send out to couple of different AWS Partners who provide payment processing services. After a few introductory conversations, the team finds a partner who they believe is technically well-poised to take this important task off their plate.

With some of these agenda item out of the way, the team could finally focus on other technical decisions that will help them to sustain their expected growth. AWS Amplify served them well during the beta stage: It helped them a lot with preparing user interfaces suitable for mobile devices. They decide to continue relying on it for building and maintaining all of their current and future front-end applications. On the backend, they want to have more control over how they build their application services and the persistence layers they rely on. With the expectation of dealing with much larger volumes of data and to prepare for the new features they are planning, the team decides to take the advice of the AWS solutions architect and start looking into some purpose-built databases. Amazon DynamoDB did great, but with the long term plans of increasing the frequency of processing market data and calculating portfolios more often they start looking at time series databases like Amazon Timestream and some relational databases like Amazon RDS for PostgreSQL. These purpose-built database services will allow the team to use the database engine that is best-suited to their different workloads.

On the application development side, the team wants to start implementing more complicated business logic without having to worry about increased operational overheads. They know they wanted to containerize their workloads but aren’t certain about which option will best fit their small team. The AWS team earns Example Startup’s trust and becomes a frequent participant in the brainstorming sessions and decision-making process. AWS’ recommendation on the container orchestration is Amazon ECS with capacity provided by AWS Fargate – the serverless compute for containers. The appeal of Fargate is that it provides a flexible scaling approach because of its pay-per-use functionality, without having to worry about patching the underlying operating system. Given the lack of certainty around the start date for the next cohort, this is a welcome option that gives the team more time to focus on their development activities.

Security is another topic gaining prominence on Example Startup’s list of priorities. With the payment solution buildout underway, the platform will include a higher risk exposure. As part of the continuous efforts of anticipating startup needs and meeting them in a proactive way, AWS has recently published the AWS Startup Security Baseline (AWS SSB) document. AWS SSB is a set of controls that create a minimum foundation for businesses to build securely on AWS without hindering agility. The team had some of their work cut out for them.

The current architecture diagram for Example Startup.

Optimizing for cloud costs with AWS

The team is busy experimenting with ideas, implementing new technology, and learning how to use the services and features they might need. With AWS Budgets already setup, the technical founder decides to get familiar with more tools to give them better oversight and control over their AWS spend. She learns about tools like AWS Cost Anomaly Detection, an automated cost anomaly detector and root cause analysis with built-in machine learning (ML) and alerts. Diving deeper into the details, she learns about AWS Cost Explorer, a tool that provides the ability to view and analyze costs and usage details.

Raising capital to support the startup’s growth

The AWS Activate credits helped with the AWS cost, but the team is growing and other expenses start piling up as well. The initial bootstrap funds are near depletion, gradually limiting the team’s ability to experiment. It is time to start thinking about raising some capital. The founders have been getting ready for this moment for some time, with a deck almost ready. This is not something they have prior experience with, nor the contacts that would be able to help. They do have AWS on their side. The AWS team facilitates conversations with the Business Development teams, who are happy to help with advice and introductions to investors and venture capital firms. Exciting times are ahead.

Check out the first blog in the Evolutionary Architectures series.

Building Black entrepreneurship with the support of AWS employees

AWS Editorial Team — Mon, 27 Feb 2023 21:06:34 +0000

To celebrate Black History Month, AWS Startups is featuring posts throughout February highlighting the contributions of Black builders and leaders in tech. Above all, these individuals inspire, empower, and encourage others—especially those historically underrepresented in tech—to prove what’s possible.

The most successful teams reflect the general makeup of society. Research shows that diversity is good for business: it leads to teams that make better, faster decisions, as well as higher employee satisfaction and higher financial performance.

Yet Black individuals are woefully underrepresented in tech: as founders, engineers, and in key leadership roles. Black adults comprise 12% of the US workforce, but only make up 7% of workers in computer occupations. It is astounding that, as of 2021, funding to Black entrepreneurs represented just 1.2% of US venture dollars (with only 0.34% going to Black women), while 14% of the US population is Black.

In celebration of Black History Month, it is crucial to shed light on the Black innovators in tech who are building and reconstructing an industry that underrepresents them, while also inspiring the next generation of Black innovators who will follow in their footsteps.

Step by step, we want to narrow the numbers gap and make the playing field more equal for everyone, and we must start by spotlighting Black innovators who are breaking barriers in the tech field. Here are a few ways AWS and its employees are working to achieve this.

Opening doors to accelerate success

Every startup faces challenges, and some face more challenges than others. Accelerators are a tool for founders that concentrate years of business experience into key learnings that founders can use to steer their business to success. The AWS Impact Accelerator Program is one way that Amazon Web Services (AWS) worked to provide equal footing among startups in 2022. The Impact Accelerator provides unparalleled resources to underrepresented founders, such as the coaching, investment, networking, and media reach needed to accelerate their business.

“It’s an enormous privilege to be able to put Amazon’s connections and capabilities to work on behalf of these amazing startups,” says Denise Quashie, Head of Worldwide Startup Marketing Programs at AWS. “If we can open just one door for each one of these startups, it really puts them on more equal footing with those who are considered the majority and have those doors open much longer.”

Recognizing Black AWS employees whose work supports Black founders

To celebrate the voices of Black employees at AWS and amplify their work’s impact, we are honored to share the stories of eight AWS employees working to advance Black entrepreneurs in ways that strengthen Black businesses and support economic growth in historically underserved Black communities. These include programs such as Impact Accelerator, REACH, and Amazon Catalytic Capital. They’re also sharing their advice for others who hope to follow in their footsteps.

Brandon Middleton, Account Manager

Brandon Middleton is an Account Manager with the AWS Fintech and Web3 startups sales team based in Palo Alto, California. He enjoys working with the builders defining the next generation of value on the internet that we call Web3. Thinking big, learning, and being curious are staples of every conversation and interaction he has with founders who leverage blockchain technology to redesign finance and digital ownership at scale. His specific focus is on supporting AWS Startups customers between the Seed and Series-C stages who are building in the Web3, crypto, and blockchain space.

What advice does he have for Black innovators in tech?

“Show up as your genuine and authentic self. Your voice is needed as a builder of this Web3 future as much as it is needed as a consumer of the products and services it will produce. The more points of view we contribute during these formative design years, the smoother the plane ride will be in the years to come, as we will have mitigated risk, shed light on vulnerabilities, and improved the overall quality of the tools future generations globally will be using daily.”

Charlotte Newman, Global Head of Underrepresented Founder and Investor Startup BD

Charlotte Newman is the Global Head of Underrepresented Founder and Investor Startup Business Development. She is located in Washington, DC, where she leads a team with a laser-like focus on how to accelerate underrepresented founders and investors. As a former founder, Charlotte says it is a privilege to use what she learned to enable other entrepreneurs to launch and scale their best ideas. She believes that what we do, at scale, has the power to democratize entrepreneurship and strengthen the role of historically-marginalized groups in entrepreneurial ecosystems around the world.

What advice does she have for Black innovators in tech?

“Throughout my career, I have found that achieving business objectives while driving positive change for historically-underserved groups at scale requires a combination of hard and soft skills. In leading a global set of programs and partnerships that advance Black entrepreneurs, I employ data-driven problem solving as well as curiosity, empathy, and humility. Ultimately, this ensures that the Black entrepreneurs served by my work achieve better outcomes and feel seen.”

Daniel Omachonu, Account Manager

As an Account Manager for Fintech startups sales in Brooklyn, NY, Daniel enjoys working with visionary startups disrupting various industries. Amazon and AWS radically disrupted retail and tech infrastructure; startups are disrupting the future of technology and everyday life, so it’s great to be a part of their journey while at AWS. Daniel’s focus is on fintech startups transforming the financial services industry.

What advice does he have for Black innovators in tech?

“I have had the pleasure of meeting some incredible Black founders with great companies and ideas. AWS is in a position to improve its access to funding opportunities and support. My advice would be that there is room for everyone to help somehow, so raise your hand, show up, and get involved.”

Dottie T, Senior Account Manager

Dottie T is a Senior Account Manager with the AWS Greenfield Early Startup Account Team. He enjoys working with early stage founders who leverage the AWS platform to solve complex issues and advance the civilization of mankind. His Focus area is Pre- Seed, Seed, and Series A-funded startups.

What advice does he have for Black innovators in tech?

“AWS is beyond a technology platform. It’s also a platform for positive change—and every employee can be part of this change, either by volunteering at the Black impact accelerator, being a soundboard for underrepresented founders, or being an internal voice of conscience that helps AWS build more programs and processes to support Black founders. We are all capable of being a source of inspiration.”

Jarman Hauser, Global Tech Business Leader

Jarman Hauser is a Global Tech Business Leader in Seattle who likes to tinker and explore unique approaches to solving complex, multi-dimensional problems. Jarman finds ways to use tech for good by leveraging the innovation at Amazon to better communities, environments, and our society. He is constantly impressed and equally inspired by the new generations of startup founders solving the world’s most complex problems. Jarman will continuously focus on equity, investing in underrepresented entrepreneurs, and creating “longer tables.”

What advice does he have for Black innovators in tech?

“Inclusion is a virtuous cycle. Continuous innovation depends on diversity, but more often than not, underrepresented founders lack the same levels of access and opportunities. We’re starting to see change, but in order to build equitable entrepreneurial ecosystems, we have to collectively take active approaches in breaking down the systemic barriers that perpetuate inequality.”

Nehemiah Green, Global Business Development Partnerships Manager

Nehemiah Green is a Global Business Development Partnerships Manager in Washington, D.C., who feels privileged to meet and build relationships with incredibly founders, every day, as they leverage technology to change the world. He finds it invigorating to learn from them and to help identify solutions to their biggest challenges. Nehemiah will continue to focus on helping under-represented founders and investors grow and succeed at different stages throughout their lifecycle.

What advice does he have for Black innovators in tech?

“After reading Bryan Stevenson’s book, Just Mercy: A Story of Justice and Redemption, a few years ago, I’m a big advocate of getting proximate on issues of systemic racism and bias. In my conversations with Black founders, I hear their stories and listen to their experiences which has been critical to my ability to support and problem solve. Being proximate builds our capacity to build solutions that are driven by empathy.”

Sekai Ndemanga, Principal Fintech Business Development

As a Fintech Business Development lead in New York, Sekai is constantly learning in her role; Fintech is broad and covers various industries. In her day-to-day work, she speaks to customers and founders influencing the future of financial services, be it the fascinating world of insurtech or proptech. She’s also the host of the popular Fintech in the Cloud podcast, and a key contributor to the 2022 State of the Industry: African Fintech report.

What advice does she have for Black innovators in tech?

“It does not have to be a grandiose initiative; start small. If you are cognizant of underrepresentation in your space, think of small ways to make a difference. Fintech can be an exclusive community with representation that reflects systematic structures not built for people of color. Therefore, we must showcase founders that are advancing against the odds.”

Tiffany Johnson, Global Business Development Manager, URFs Program

Tiffany Johnson, of Seattle, Washington, is a Global Business Development Manager on the Underrepresented Founder/Investor Business Development Team. She finds fulfillment in connecting with founders and understanding their needs as they build their businesses through engagement such as workshops, summits, dinners, and more. Listening to the customer’s voice is a key aspect of her work. Tiffany’s area of focus is on creating global initiatives to support startup founders from underprivileged backgrounds.

What advice does she have for Black innovators in tech?

“My advice to others who hope to make a positive impact in the Black community through their work is to first, understand the unique challenges and barriers that Black entrepreneurs face, and second, to actively seek out and build partnerships with organizations and individuals who are already working to address those issues. Only by working together can we truly drive meaningful change.”

Black founders building successful startups on AWS

We hope that, through our platform, we can provide an innovative and inclusive space to underrepresented communities that will inspire the next generation. AWS is committed to supporting equitable and inclusive access to building on the cloud.

Explore more content that celebrates the achievements of Black innovators, such as:

More about the AWS Impact Accelerator program

The AWS Impact Accelerator Program opened the doors to creating access and resources for underrepresented leaders and founders in tech. If you are interested in applying for the upcoming AWS Impact Accelerator: Latino Founders (or if you know a great candidate), learn more here.

Why culture is critical to startup organizational growth

Hannah Wall — Thu, 23 Feb 2023 19:23:28 +0000

Startups are in the business of proving what’s possible and bringing world-changing ideas to life. But founders can’t do it all on their own. To achieve their goals, they must build strong teams they can trust to implement their vision. And they need to cultivate a strong workplace culture that encourages experimentation, quick decision making, and learning from mistakes.

Indeed, new research commissioned by AWS shows that workplace culture is vital to a startup’s success. Six out of seven startup leaders (86%) believe a company’s culture contributes to its growth, with 85% saying it can play a critical role in securing investment and the same number saying it is an important factor in attracting new talent.

To build a strong workplace culture, leaders must do more than offer “traditional” startup perks like ping-pong tables, kombucha kegs, and paid lunches. Craig Cowdrey, co-founder and CEO of the workplace wellbeing startup Sonder, says today’s workers have become skeptical of these basic benefits. “So much of these [traditional perks] have been recognized as not particularly relevant or determinative in an employee’s choice,” he says. “A lot of them think that it’s just to keep them working in the office longer.”

Instead, the research indicates that startups should prioritize work-life balance initiatives in order to attract talent and get the most out of their teams. Startups can often be chaotic, high-pressure environments, but founders must work to protect both themselves and their teams from the “grind culture” of overwork and burnout that is so prevalent in the startup ecosystem.

Dismantling that culture of constant work and pressure will not happen overnight — 93% of startup leaders we surveyed acknowledge that grind culture currently exists in the startup landscape. But a growing number of leaders recognize that logging long hours in the pursuit of perfection or to fend off the competition isn’t a recipe for success in the long term.

“Pulling 60- or 70- or 80-hour work weeks isn’t a badge of honor — it’s stupid, and you can’t keep doing that,” says Megan Woff, head of founders at Startmate Accelerator.

Instead, developing a workplace culture that emphasizes self-care is key to a startup’s growth and long-term viability. Dr. Ben Hurst, founder and CEO of the patient engagement platform HotDoc, says that as his company has evolved, they have worked to minimize the grind and embrace wellbeing as a priority. “You are doing the wrong thing if you are working in a way that is not self-sustaining,” he says.

Hurst’s way of thinking is likely to have ripple effects across the business landscape, working to create healthier workplaces and putting companies of all sizes on the path to success. 71% startup leaders say the larger corporate world watches the startup landscape very closely. And one-third of startup leaders believe that by disrupting older, outmoded ways of working, startups are having a positive impact on the wider business culture.

AWS is proud to support startup founders on their journeys, helping them to solve the world’s problems through the power of technology. The AWS Activate program provides qualified startups with a host of benefits, including AWS credits, technical support, and training.

How Resilia is helping nonprofits to build capacity through SaaS

Sevetri Wilson — Wed, 22 Feb 2023 13:58:40 +0000

To celebrate Black History Month, AWS Startups is featuring posts throughout February highlighting the contributions of Black builders and leaders in tech. Above all, these individuals inspire, empower, and encourage others—especially those historically underrepresented in tech—to prove what’s possible.

Sevetri Wilson, founder and chief executive officer (CEO) of Resilia

The AWS Startups Blog is excited to introduce Sevetri Wilson. Sevetri is a serial entrepreneur whose most recent startup, Resilia, enables nonprofits to increase capacity and funders to go beyond the grant with technical assistance, coaching, and capacity-building support. In October of 2022, Resilia closed a Series B $35M funding round, the largest raise ever for a solo Black female founded tech company.

“Resilia is proving that technology can break through the perceived number of finite resources. When you’re a consultant, there’s a finite number of people you can help and a finite amount of time in which to do it. By delivering resources through technology, Resilia is using tech for good: To revolutionize the way leaders develop and grow their organizations and to remove limitations on the things they oftentimes have no say in.”

Can you please introduce yourself, as a person and as a professional?

As a person, I am—as I always say—a girl raised in the South. I’m from Louisiana and was born and raised in a small city about 45 minutes outside of New Orleans, called Hammond, where my mother was raised. I was raised by a very big family: my mother was one of nine and my father was one of seven, so you can imagine how many cousins and family members I was fortunate enough to grow up around. My grandfather was a farmer, so we had blackberry bushes and a sugar cane field and cows and chickens running around.

Professionally, I would consider myself a serial entrepreneur and business owner. I’ve bootstrapped one company. I’ve now raised close to $50 million for Resilia, a technology company that I founded in 2017. As a professional, I’ve grown into a space where I am known as a problem solver. People always ask, “What do you think has led to your success?” I say that as a professional I’ve always been able to solve people’s problems.

What is the founding story of Resilia and the company’s mission?

I founded my first company, Solid Ground Innovations, in 2009. We were a management and consultancy agency and we had a nonprofit arm called SGI Cares. We worked as consultants to drive strategic functions around community giving with large funders like the W.K. Kellogg Foundation and family foundations, and large corporations like Aetna Better Health and Community Coffee. We would come in and work with them to deploy resources and capacity support to the initiatives that they were funding from a philanthropic standpoint.

Although we would come in and we would bring capacity, the likelihood of that work continuing once we left was very rare. I started thinking about ways that we could productize our services and deliver them through a software solution to create a continuum of service that didn’t stop once we left.

That’s what gave rise to Resilia. We are a two-sided platform: On one side, we support nonprofit organizations, helping them bring capacity to their day-to-day through our software platform. On the other side, we enable large funders such as private foundations, public charities, corporations, and government entities to deliver capacity-building resources at scale to the nonprofits they support.

A lot of Resilia’s journey was me taking something that was heavily based in consultancy and productizing it, to bring a more digitized presence to the work through software. Resilia digitizes the work and democratizes philanthropy to make what generally only a few nonprofits would receive—whether that’s resources or something else—more accessible to everyone.

Today, we’re at over 100 employees and primarily based in New Orleans, along with an office in New York, an engineering office in Mexico, and remote employees across the United States.

What are you most proud of accomplishing as the founder of Resilia?

My proudest achievement is growing a very diverse team around tech for good. We have people of color and women in every single rank of Resilia: From the CEO, of course, to our VPs, to our directors, managers, and our entry-level team members.

Resilia’s platform for non-profits is radically and successfully transforming how they do business. Can you share some of the opportunities and challenges for startups that are bringing technology to the non-profit industry?

Historically, I would say that technology has not been built with nonprofits in mind, and that’s probably why sometimes our space has been averse to adopting technology. That’s a challenge that we are faced with: How do we build the most user-friendly product that’s for nonprofits as a whole? We believe we’re building Resilia as the new age of technology to help nonprofits have the resources to keep going, so they can do their work.

Has AWS supported your goals for Resilia?

Yes. We build a lot of our products on AWS and AWS solutions are an important part of our stack.

Can you share the most important lesson you’ve learned as a founder?

Deliver solutions that increase efficiency and effectiveness at scale. For instance, with the Resilia platform, nonprofits can streamline their operations, reduce administrative burdens, and focus more on their mission-driven work.

In October 2022, Resilia closed a Series B $35M funding round. This is the largest raise ever for a solo Black female founded tech company. Do you have any advice for other founders about how to succeed at funding rounds?

Founders who are building right now—similarly to small businesses—are seeing market volatility. You have to be really scrappy and stay as lean as possible to survive what we are seeing as a drawback in funding from investor groups and funders alike.

Also, when you’re going out to fundraise, ensure that you have all of your i’s dotted and t’s crossed, meaning have all of what you need for due diligence: have the financials together, have reference calls and who’s going to do those reference calls. Have as much done as you can to limit the time you have to be out in the market raising. Ensure that you run a very tight and smooth process so that you aren’t putting yourself in a position where you are running out of capital. The most that you can do is just be prepared and have, as I would say, your house in order so that you save time.

What’s next for you and Resilia?

In 2023, Resilia will continue to scale our team in order to serve our growing client base. We’re also really excited to roll out donations and payments. That’s a new feature that we’ll be offering to non-profits and our existing customer base as a whole. We want to continue to expand our offerings so we can truly be a one-stop resource for nonprofits.

We’re also rolling out a robust online community to foster more peer-to-peer learning. We’re using this Resilia community to provide a guided course track to help nonprofit leaders not only build knowledge, but also upskill their teams. This year we’re also going to deliver more impactful features and products to funders and grant makers alike.

Explore more content that celebrates the achievements of Black innovators, such as:

AWS Startups Twitter stories of inspiring, innovative, and game-changing Black founders
CareCoPilot founder Alyse Dunn wins big after AWS Impact Accelerator
Shining a light on Black excellence: three founders share their stories
How Eyegage scaled their life-saving app via Impact Accelerator partnership

Fighting funding disparity with Amazon Catalytic Capital

Howard Wright — Tue, 21 Feb 2023 15:33:51 +0000

In October 2022, Amazon launched its Catalytic Capital initiative, investing $150 million in funds to underrepresented entrepreneurs. Now, AWS is further accelerating the conversation through strategic partnerships and data-driven reporting with the publication of the State of Black Venture Report, launched Tuesday at the 2023 State of Black Venture event hosted by BLCK VC at the AWS Startup Loft in San Francisco, California.

When new founders have access to venture funding and mentorship, they go on to build solutions that transform our society, create jobs that strengthen our communities, and inspire the next generation of founders to begin their journey toward entrepreneurship and innovation. Unfortunately, not all entrepreneurs and startups are funded equally.

While the funding disparity is well-documented, not many people understand the full scope of the issue. For instance, of all available venture funding, underrepresented founders receive just 1.87%. Black, Latino, women, Indigenous, and LGBTQIA+ entrepreneurs miss out on more than 98% of the financial backing that could take their startups to the next level.

What’s more: funders of marginalized identities are underrepresented in the venture capital industry. According to State of Black Venture Report, a 2022 report published by BLCK VC, Black partners in 2020 comprised just 3% of all partners across the industry. The same report tells us that Black funders are up to four times as likely to fund Black startups than non-Black funders.

Let’s be clear: Black entrepreneurs are underrepresented and underfunded because the people most likely to invest—Black funders—are underrepresented and underfunded. The result is a lack of the kind of resources that allow startups to grow and scale. It also creates a vacuum in the mentorship that propels both entrepreneurs and junior investors toward success in their earliest stages.

The same is true for and women founders and funders. While Latino investment professionals grew in 2022 according to LatinxVC, they still only represent 2% of the overall venture funding industry. And only 35% of those professionals identify as women. Minority- and women-led firms manage just 1.4% of the $8 trillion venture capital and private equity industries.

The data is clear: Diversity in entrepreneurship relies on diversity among investors.

AWS has taken a collaborative and multifaceted approach to the issue. Initiatives like the Amazon Catalytic Capital initiative and the AWS Impact Accelerator deploy capital directly to underrepresented founders. Not only that, but AWS is funding organizations and research that bring the venture capital disparity into clearer focus and help founders and funders connect.

The role of AWS is to foster an ecosystem that facilitates a network that brings underrepresented founders, funders, and the organizations into closer contact so that they can support one another. AWS has teamed up with organizations like Diversity VC, LatinxVC, StartOut, and BLCK VC, whose missions are not only to support underrepresented founders and funders, but to produce data-driven reports that lend credence to the conversation.

Howard Wright, Sydney Sykes, Toussaint Bailey, and Jessica Murrey at the State of Black Venture Report discussion panel.

This month, BLCK VC hosts its second annual State of Black Venture event on February 22nd at the AWS Startups Loft in San Francisco. The organization convenes a panel of Black leaders to share their triumphs and challenges on the road to becoming funders. BLCK VC’s accompanying report, State of Black Venture, gathers data through surveys, interviews, and extensive industry research to elucidate the sobering issue of the funding disparity.

The report also offers some encouraging news:

The number of Black partners is growing. Through their research, BLCK VC discovered that 83% of Black investors have either launched their own fund or have joined new firms where they have greater influence. Nearly a third (27%) of these funds launched in the last two years, signaling a strong upward trend. What’s more, more than half of all Black funders are reported to be actively mentoring Black junior mentors, either within their network or at their firm.

Digitalundivided finds that, despite the reality of the venture funding gap, underrepresented founders continue to persevere. Their 2022 Industry Insights Report reveals a positive trend in leadership among Black and Latina women founders, especially in healthcare, financial services, and education. Not only are Black and Latina women more likely than ever to found startups in these industries, but they are more likely to secure funding.

There’s more work to be done, and shedding light on the funding gap through impactful, in-depth research is a vital part of the process. The better we understand our challenge, the better we can work towards a solution. Publications such as the 2022 State of Black Venture Report acts as a guiding light as we work with our partners to build pathways to success for underrepresented founders.

Learn more about Amazon’s Catalytic Capital initiative and the AWS Impact Accelerator.

You may also want to join the upcoming AWS Impact Accelerator: Latino Founders. Applications are open from March 6 – March 17, 2023.

Shining a light on Black excellence: three founders share their stories

Tiffany Johnson — Wed, 15 Feb 2023 17:14:04 +0000

To celebrate Black History Month, AWS Startups is featuring posts throughout February highlighting the contributions of Black builders and leaders in tech. Above all, these individuals inspire, empower, and encourage others—especially those historically underrepresented in tech—to prove what’s possible.

Today, we’re talking to three Black founders and leaders: Kwame Boler, CEO and co-founder of Spritz Natasha Greene, We Intervene; and Chandler Malone, CEO of Bootup.

These Black leaders and innovators are making impacts in their communities, industries, and beyond.

Read on to see how they’ve overcome obstacles and how they encourage and mentor young Black founders in tech and beyond.

Kwame Boler is the CEO and co-founder of Spritz, the full-service app for residential cleaners to handle paperwork, stay organized, and earn more.

Q: Tell me about some of the obstacles you’ve faced as a founder. What did you learn about yourself as you confronted them?

A: As a founder, I’ve faced many challenges along the way. One of the biggest obstacles I experienced was when my previous business, neu, had to shut down and pivot to become Spritz. At the time, it was a low point for our team, but we didn’t let it defeat us.

We started neu with just $3,000 of funding from friends, and by the end of 2019, we bootstrapped the business to generate nearly $400,000 in gross merchandise value (GMV) in just one city. In 2020, our goal was to close our first round of venture capital while participating in the Techstars accelerator and expand to other markets. But, unfortunately, when the travel industry took a nosedive from the pandemic, our plans had to change. However, we saw this as an opportunity to clear our technical debt, and after making the tough decision to furlough our operations team, we continued building and increasing our product-market fit by including personal protection equipment (PPE) in our service offerings. Despite the circumstances, we successfully closed our first round of venture capital that year. We were now on our way to scale, and by the end of 2021, we had doubled our lifetime GMV to nearly $900,000 in sales and expanded into our 2nd market.

However, while raising our next venture round, our most significant challenge came when our lead investor failed to wire funds at the last minute despite several reassurances and a binding commitment.

Consequently, we had to lay off 90% of our team and reevaluate our plan. It was a tough time, but in 2022, we eventually pivoted to Spritz, a SaaS platform that simplifies the complexity of managing the back office for residential cleaners. We spent most of last year doing customer discovery and market research, and by October, we were pleasantly surprised to see nearly 4,000 people had signed up for our waitlist in anticipation of our product launch. Despite our hardships, I am incredibly grateful for the team that helped us get to where we are today.

These experiences taught me that sometimes, the hardest problems could be blessings in disguise.

Q: What advice do you have for young founders who want to start their own company?

A: For those ready to embark on the exciting journey of starting your own company, here are the top 5 lessons I’ve learned along the way:

Despite the challenges, bear markets can be the best time to build a company.

My top tip for building a successful team is to leverage tools like LinkedIn to source talent and surround yourself with people you know, love, and trust, who have complementary expertise to yours. Then, take a good hard look at your strengths and weaknesses, and seek out individuals who can complement those and bring different skills to the table. Building a strong, cohesive team is the foundation of a successful company.
Next, please don’t quit your day job (… yet). Moonlighting can be a game-changer and an excellent opportunity to test your entrepreneurial skills, validate your business idea, build a savings nest egg, and prepare for success. By dedicating your evenings to your passion project, you’ll stay focused and motivated and won’t lose your day job security. This way, when the time comes to quit your day job, you can fully concentrate on your startup without worrying about money.
Then, before you dive into building your product, it’s crucial that you spend time researching the market and validating your assumptions with customer feedback. This will help avoid blind spots and ensure you’re building with your customers in mind.
Also, understand that your net worth is your network. Building a successful startup takes a community, just like raising a child takes a village. Put yourself and your idea out there by attending as many startup networking events as possible. Build relationships and surround yourself with like-minded individuals who have experience and can give you the right advice. Leverage a network of advisors and potential investors to help you filter the noise and make efficient decisions.
Finally, please remember to prioritize self-care as well. Starting a company is a marathon, not a sprint, and it’s essential to pace yourself and be kind to yourself and your team. It’s okay to take a step back and celebrate your wins — just be sure to maintain open lines of communication with your team and investors.

With a lot of hard work, determination, and a little luck, you can also build a successful company.

Natasha Greene is the founder and CEO of We Intervene, which provides a centralized repository of resources along with Real-Time Virtual Resource Assistants to better help schools connect families to resources they need and want 24/7.

Q: Tell me about some of the obstacles you’ve faced as a founder. What did you learn about yourself as you confronted them?

A: I delivered a stillborn baby on December 1st. Her name was Jurni. My stillbirth was a big emotional and physical setback as a woman in tech who wants to have a family. I wrote about it on LinkedIn.

I learned that community could help me to get through almost anything. For example, I could continue building out the security requirements for We Intervene because of Audree, whom I met in AWS Reach. She became my part-time virtual chief information security officer (VCISO) and helped oversee development and milestones to move our product through development for a big meeting we had in February while I work on healing over the death of my daughter.

Staying patient is also an obstacle. Some sectors are slow to realize the potential of your product, and you have to keep on educating others on the impact your product will have. And you have to keep on reminding yourself why you decided to take this entrepreneurial journey and remember the impact you want to have. Many people see the “quick” successes, but it wasn’t quick at all; there were many days between the startup phase and the first purchase stage. There are many more days between the startup phase and the exit or even grossing millions of dollars phase. The key is always patience – patience with yourself and all those around you.

Q: What advice do you have for young founders who want to start their own company?

A: You need to go out there and start the project/business you’re thinking about doing. But first, start by testing your assumptions on the idea you have. For example, if you want to start a TikTok marketing company, then you need to do it for free for a couple of people and get their feedback on your work.

Or, your “free” time can be tons of projects you have completed. Then, you take your positive testimonies or reviews to the next group, who will pay. And then there is another group that will pay you more as you become better at your craft. Always work on getting to clarity on what you are offering and who you are offering to – we don’t want to say you are for everyone when you really want to focus on solopreneurs.

Also, know that the first company might not be the thing you run with. It might take a couple of companies. So, you might talk to an accountant about the best business structure you should do to house all your ideas.

If I had to do entrepreneurship again, I would have done an S-corp in the beginning and then d.b.a. all my business ideas under that idea instead of starting an LLC for each idea I have. I wasted a lot of money on business structure and taxes.

Chandler Malone is a three-time entrepreneur who is now building Bootup, an educational labor marketplace that helps individuals get their first jobs in the technology sector regardless of their educational background, while helping companies fill their talent pipeline problems through access to pre-qualified talent.

Q: Tell me about some of the obstacles you’ve faced as a founder. What did you learn about yourself as you confronted them?

A: One of the most difficult things that no one tells founders about is building and growing a strong team. As a startup, the capital we raised definitely gave us flexibility, but we still can’t compete with big tech companies on salary, benefits, and resources.

One of my greatest strengths is my ability to be resourceful and think out of the box for solutions. This resulted in building out a fully distributed, in house engineering team across Bolivia and Venezuela. We were able to hire the most experienced engineers and within our budget.

In 2023 and onward, managers should be focused on understanding the best ways to source and ration top talent globally. We are no longer confined to specific geographies teams who can find and keep the best people no matter where they are located will succeed in the long run.

Building our team has also pushed me to become a stronger listener, more perceptive, and trust my gut more.

Q: What advice do you have for young founders who want to start their own company?

Don’t get caught up in the hype and stay focused on running your race. You’ll see some founders always in the press or out at conferences – and good for them!

Your job isn’t to emulate what other founders are doing or get the validation of VCs and the press. At the end of the day you are running a business, and anything that isn’t focused on better serving customers is a distraction.

Conclusion

We’ll let Kwame take us out with a final question.

Q: What makes you excited about or gives you hope for the future of Black-owned and led startups?

A: The incredible progress I’ve seen over recent years gives me hope for the future of Black-owned and led startups. We’re now seeing more and more Black founders getting access to resources and capital and a growing community of supporters helping to blaze the trail for others to follow.

The ecosystem is evolving, and with a combination of intentional actions and continued success stories, I’m confident we’re moving toward a brighter future for Black entrepreneurs. Of course, systemic issues won’t change overnight, but it’s inspiring to see the progress already made.

I’m particularly excited about the rise in Black-led funds, which reduce the perceived risk of investing in Black founders. And with access to wealth increasing among Black people, we’re going to see even more Black entrepreneurs taking their ideas to the next level.

In the tech community, we often expect change to happen as quickly as writing a line of code. But we’re learning that some issues are systemic and deeply rooted. However, the progress being made and the quality of the founders getting access to resources gives me confidence that, over time, more founders of color will get the opportunities they deserve.

I’m hopeful and confident about the future of Black-owned and led startups. I can’t wait to see more Black leaders trailblazing in this industry and giving back to help others.

14 reasons to fall in love with AWS Activate

Megan Crowley — Tue, 14 Feb 2023 14:59:34 +0000

Got a great idea? Build it with AWS Activate. AWS Activate is our free startup program that provides tools and resources including AWS credits, AWS Support credits, exclusive member-only offers, personalized guidance, expert advice and more. It is designed to help startups build, launch, and grow.

Why AWS Activate for your startup?

AWS Activate is your solution to a scalable, reliable, and cost-optimized startup:

Get started on AWS at no cost with up to $100,000 in AWS credits.
Get technical support and architecture guidance with up to $10,000 in AWS support credits.
Build your infrastructure in minutes with pre-built templates using best practices.
Take advantage of special offers including discounts, free products, memberships, services and tools.
Get access to the Activate Console that provides personalized guidance, details of your AWS credits, and more!

AWS Activate members get to choose which Activate tier best suits their startup journey: AWS Activate Founders is for early stage startups (unfunded and funded up to and including Series A). AWS Activate Portfolio is for startups affiliated with an Activate Provider.

AWS Activate Offers

To help you show your startup some love on Valentine’s Day, we’ve curated a list of exclusive offers from AWS Activate Providers ?.

Airtable

Airtable enables any team, regardless of technical skill, to create apps on top of shared data and power their most critical and unique workflows. Teams at more than 300,000 organizations, including 80% of the Fortune 100, rely on the Airtable Connected Apps Platform to stay aligned, execute with greater agility, and connect previously siloed teams and data.

Get the offer: $2000 in Airtable credits and access to the Airtable for Startups program. One credit redemption per company.

Brex

Startups can scale faster with a Brex corporate card and business account. It’s the smartest, fastest way to deposit funds, send no-fee payments, and effortlessly track expenses. Get 10-20x higher limits, no personal guarantee, and tailored rewards on every purchase.

Get the offer: AWS Activate companies get 80,000 points after spending $10,000.

Carta

Startups get started for free (for startups with $1M raised and/or fewer than 25 shareholders), and when you upgrade to a paid plan, you’ll receive 20% off your first year subscription. Receive cap table management, fundraise benchmarking, SAFE tracking, and more.

Get the offer: 20% off a Carta paid plan.

Deel

Hire internationally with complete confidence. Deel will handle your worldwide compliance, payroll, and HR in 150+ countries, so you can hire the best talent regardless of location. Get 20% off the first year when hiring full-time employees and contractors.

Get the offer: Free Global HRIS and 20% off Compliance and Global Payroll for one year (a value of up to $25,000!).

Firstbase

Our seamless incorporation service is built with you in mind. We’ll handle the admin work, while you focus on building your business. Access $150,000 worth of exclusive rewards from Brex, AWS, Airtable, G Suite and more.

Get the offer: A 10% discount on US incorporation.

Pipedrive

Pipedrive is the first CRM designed by salespeople, for salespeople. Do more to grow your business.

Get the offer: Try it free for 30 days, plus get 50% off for 1 year.

DocSend

Dropbox lets you save, access, and share all your work in one organized place. DocSend’s document analytics show who is looking at your pitch deck and which pages capture attention. Deliver an intuitive investor experience with HelloSign’s easy-to-use eSignature solution.

Get the offer: Get up to 90% off DocSend, 50% off your annual HelloSign Essentials or Standard plans, & 40% off annual subscriptions of Dropbox Business Standard and Advanced.

Mercury

Mercury has intuitive dashboards and payment flows, virtual and physical debit cards, team management capabilities, and API access. Setup FDIC insured checking and savings accounts with no monthly fees, use treasury products to save and grow your money, use their foreign currency exchange features, find your next funding round through their Raise program, or take out a line of credit through their Venture Debt program.

Get the offer: $500 cash for your startup if you deposit $50K within 90 days of approval; $250K in accounting services, software and other perks; Free domestic and international wires & ACH (in USD); Free banking and financial planning stack; Access to Raise and First Check

Want to show your startup the love all year round?

Check out the AWS Activate Console for more exclusive offers for AWS Activate members. The AWS Activate Console is your personalized hub of tools, resources and content tailored to your startup needs. Designed to support you through every stage of your startup journey, from ideating to building and beyond, it’s a one-stop-shop that delivers the tailored solutions you need to quickly get started on AWS and grow your business.

Want to share an AWS Activate exclusive offer from your startup? Learn how to become an AWS Activate Provider today.

How Eyegage scaled their life-saving app via Impact Accelerator partnership

Bonnie McClure — Wed, 08 Feb 2023 14:39:05 +0000

To celebrate Black History Month, AWS Startups is featuring posts throughout February highlighting the contributions of Black builders and leaders in tech. Above all, these individuals inspire, empower, and encourage others—especially those historically underrepresented in tech—to prove what’s possible.

KEY RESULTS

Exposes mobile app eye scans to backend
machine learning models and returns results
Allows app to perform for investor demos
Enables app to scale when workload spikes occur
Improves the security posture of the app
Sets the stage for the app to go to market

STARTUP INFO.

Dr. LaVonda Brown, founder and CEO of Eyegage, whose expertise lies in artificial intelligence (AI) and eye-analysis technologies, developed a robotic engagement model based on user eye gaze and pupil size (dilation and constriction).

This published, patented, and licensed model has been applied to several use cases, including math education and physical therapy to improve outcomes. Her work has also led to using eye tracking as a viable biomarker for mild cognitive impairment and early-onset Alzheimer’s Disease.

LaVonda used her expertise to launch EyeGage, a mobile app that applies eye-analysis techniques to evaluate whether individuals are under the influence of drugs and alcohol to help prevent fatal accidents.

Eyegage planned to go to market with their mobile application in December 2022, but they had to deliver on a key requirement: getting the frontend app connected with Amazon SageMaker—the backend cloud platform that enables app developers to create, train, and deploy machine learning models.

“Our prototype worked well with Amazon Web Services (AWS), and our model was already trained,” says LaVonda. “But we needed to expose the model to SageMaker to ensure we could scale our services as user activity spikes.” But without prior experience with SageMaker, working through the documentation proved somewhat difficult for the Eyegage team. That’s when LaVonda turned to AWS for help.

AWS facilitates partnerships via Impact Accelerator program

Via her participation in the Black Founders cohort for the AWS Impact Accelerator Program, LaVonda gained access to personalized coaching, capital funding, and technical solutions. This connected EyeGage to Avahi, a cloud-first consulting company and AWS Global Startup Program partner.

“Avahi impressed us with their knowledge about machine learning models and their understanding of our business,” says LaVonda. “More importantly, they presented previous SageMaker projects they had taken on that were similar to what we needed. That gave us confidence Avahi could do the job.”

Avahi layered the code of the machine learning model in SageMaker to expose it as a seamless API to end users of the EyeGage mobile app. This included updating the model code so it can be exposed as an event-driven machine learning inference. Avahi also layered the model with additional services like AWS Lambda for serverless computing and Amazon API Gateway, a managed service that simplifies creating and maintaining APIs.

LaVonda meeting with a coach at the AWS Startup Loft in New York to finalize her pitch for Investor Day

Scaling to save lives

With the machine learning model exposed, the frontend app could better scale to provide contactless, non-invasive, objective/unbiased, secure, accurate, and quick drug screening results.

By pursuing this partnership, Eyegage, with Avahi’s assistance, was also able to streamline the collection of end user data (such as identity and location), which allows the app to compare current and past scanning results and provide users with additional valuable information about their condition.

“Avahi also helped encode the AWS backend to receive JSON web tokens,” adds LaVonda. “This gives us a more secure way to send data back and forth, which is critical, given the sensitivity of the information we process for our customers.”

LaVonda meeting with her technical mentor in Seattle during week 1 of the Impact Accelerator for Black Founders

What’s next for Eyegage?

EyeGage is actively researching and updating its mobile application to include features to improve and assist in individual and community safety to decrease accidents. New application features, including Should I Drive? and FriendGage, promotes easy ways for accountability and accessibility to understand impairment levels.

There may also be a potential use for the company’s dataset beyond its immediate use for detecting substances in the body. “You can identify someone by their eyes or diagnose illnesses, concussions or diabetes. Or, you can tell something like if you’ve had caffeine, depending on how it responds to light.” LaVonda adds. “Monitoring eye behavior can be used for so much.”

Explore more content that celebrates the achievements of Black innovators, such as:

AWS Startups Twitter stories of inspiring, innovative, and game-changing Black founders
CareCoPilot founder Alyse Dunn wins big after AWS Impact Accelerator
Shining a light on Black excellence: three founders share their stories
How Resilia is helping nonprofits to build capacity through SaaS

Building serverless on AWS to scale Ramp’s fast-growing finance automation platform

Alexis Gordon — Tue, 07 Feb 2023 16:05:57 +0000

For startups, coming full circle is a milestone defined by partnering with the programs used during early stage growth, or providing resources that help other startups succeed as well.

Ramp, a B2B fintech startup founded in 2019 by veteran founders Eric Glyman and Karim Atiyeh, does both. Ramp is a tech-first finance automation platform whose serverless modern application–in conjunction with its corporate card–allows businesses to more efficiently manage their finances.

In the startup’s early days, founders Eric and Karim prioritized talking to customers to learn their pain points, priorities, and what aspects of a corporate card really mattered. Informed by customer needs, they tailored their product to offer:

Physical and virtual corporate cards with unlimited 1.5% cash back
Zero touch expenses to help control, analyze, and optimize organization-wide spending
Fast bill payments for businesses to pay invoices how and when they want, around the globe
Intelligent insights, reporting, and perks to maximize saving and cut spend

Within one year of launching publicly, Ramp reached unicorn status and became America’s fastest-growing corporate card. The company has since significantly scaled its business operations and AWS architecture to reach 12,000+ customers. To date, Ramp has saved businesses over $300 million and 3.5 million hours.

“The problem we solve is, ‘How can we save businesses time and money, while empowering their employees to spend, but ensuring that it’s done in a controlled and efficient way?’” explains Alexis Gordon, leader of Ramp’s product partnerships team.

Building a modern architecture on AWS

To support the startup’s need for a scalable modern architecture, high developer productivity, multi-region availability, and optimized cloud costs, Ramp built its platform’s core infrastructure on AWS.

A scalable modern architecture

“This is the modern decade of thinking about cloud infrastructure, instead of the bare bones approach to cloud computing,” explains Lewis Drummond, head of infrastructure at Ramp.

“I’m very proud of how few legacy-type virtual machines we have and that we leverage more advanced, completely serverless, technologies from AWS. It serves us very well,” says Lewis.

Ramp uses an Amazon Aurora database cluster, as well as Amazon ElastiCache for Redis to provide sub-millisecond latency for Ramp’s caching needs and to accelerate application and database performance. Jun Isaji, director of cloud infrastructure at Ramp, explains, “AWS solutions allow us to be flexible to meet demand and add components to increase system robustness. They also help us reduce complexity throughout the system by utilizing the features built into AWS solutions.”

Improved developer productivity

Ramp’s architecture uses Elastic Load Balancing (ELB), specifically Application Load Balancer, to distribute incoming application traffic. Behind that, their web servers run on Amazon Elastic Container Service (Amazon ECS) on AWS Fargate, which allows Ramp engineers to focus on building their application instead of managing their servers.

“AWS really helps by abstracting away the details of running all of our components,” explains Jun. “Our developer velocity across the organization has significantly increased from using AWS.”

Ramp also increases developer velocity by using the flexibility of AWS’ managed services to quickly and easily spin up stacks that allow them to experiment, and then spin down the stacks when they’re no longer needed.

“AWS’ managed services allow us to do proof of concepts quite easily and quickly,” explains Lewis.

“About a year ago we were looking to test Airflow, which can be a pain to set up by yourself.” To make the testing easier, Ramp leveraged Amazon Managed Workflows for Apache Airflow.

“AWS helps a long way to getting us up off the ground more quickly. Being able to go from zero to one in a matter of days instead of weeks, as well as the lower effort there, helps us to iterate quickly,” says Lewis.

Availability across multiple regions

In addition to using AWS for its high scalability and benefits to developer productivity, Ramp uses AWS’ multi-region availability. For startups, multiple regions can improve the user experience by providing low latencies across the globe and by creating more resilient cloud architecture.

Lewis explains, “These managed services within AWS work very strongly with our multi-region requirement. Having all of these managed services, which also support cross-region, has been very useful to us.” Ramp uses Amazon Aurora Global Database for cross-region with Aurora, Global Data Store in ElastiCache for cross-region with ElastiCache, AWS Secrets Manager cross-region, and Amazon S3 cross-region.

One of the most essential components of Ramp’s architecture is called the authorizer, which approves or denies credit card transactions. “Because the authorizer is so critical for us, we have a warm standby multi-region configuration,” says Jun. “We can spin up the authorizer compute within our disaster recovery region, then route requests to that compute if our primary region were to go down”

Optimizing the costs of cloud computing

Saving money on cloud spend is a priority for many startups. With the help of AWS tools and their AWS account team, Ramp has been able to decrease their cloud spend.

“Our account manager Xavier was very proactive about reaching out to us about how to reduce costs,” says Jun. “I’m definitely happy with AWS proactively reaching out and saying, ‘Here are some ways to reduce costs.’ That’s great.”

One cost-optimization success that grew from a meeting between Ramp and their account team was implementing AWS Graviton processors for Ramp’s databases. “Graviton was a big success for us in increasing performance relative to cost,” says Jun. “We’re also in the process of working with our account team to review our reserved capacity for compute.”

Tools such as AWS Cost Explorer, “make it pretty easy to understand the costs and where you might be wasting money,” Jun says. “We use AWS Cost Explorer often. It allows us to understand and trace back any big jumps or spikes in spend to a certain component or a certain change in the system.”

AWS Cost Explorer allows startups to visualize, understand, and manage AWS costs and usage over time.

Using AWS Savings Plans, which offer a flexible pricing model, “is definitely a big cost reduction as well,” Jun says.

Integrating AWS Activate into Ramp’s go to market strategy

As Ramp continues to succeed at building the next generation of finance tools, they’ve engaged with AWS Activate for each stage of their startup journey. AWS Activate is a free program specifically designed for startups that offers resources for getting started on AWS.

“Activate has helped Ramp succeed from the product perspective,” says Lewis. “The overall program has been instrumental in both Ramp’s success and also that of some of our customers.”

As Ramp grew, they joined AWS Activate Providers, a program for startup-enabling organizations to provide AWS Activate benefits to their affiliated startups. As an AWS Activate Partner, Ramp offers AWS Activate benefits to their customers, as well as a $500 sign-up offer for their product.

Alexis explains, “Through Activate Providers, we’re able to offer up to $100k of AWS credits for Ramp clients. There’s a strong overlap in our target customer base and it’s a great lever for us to deliver more time and money savings to our customers, in line with our core mission.”

Tips for developing on AWS

For developers who want to build on AWS, Lewis and Jun share some insight and best practices that serve them well at Ramp:

To gain speed, keep it simple. “Following the established patterns on AWS allows you to innovate really quickly; there’s a well-trodden path for developers wanting to start companies on AWS,” advises Jun. “In particular, I’ve had good experiences working with the solutions architects. When we have questions, they give us a lot of good insight into what’s the simplest way and how they’ve seen it work in the past.”
Harness the appropriate permissions and resource sizing from the beginning. “Six months later, when your startup is off the ground, that sets you up for success in the long run,” advises Lewis. “It helps you to pass security audits and ensure your company’s finances—and your $100k in Activate credits–last you longer.”

The future of fintech and Ramp

Ramp expects the list of fintech innovations to continue to grow: Buy-now-pay-later, embedded finance options, flexible payment terms, and revenue-based financing (to name a few) are simply the beginning.

“The emergence of fintech as an industry sparked change in a financial services sector that had been dominated by large banks for hundreds of years,” explains Alexis. “Agile, nimble, customer-focused startups like Ramp came into play to create great customer experiences and products.”

Ramp’s upcoming plans include increasing automation, streamlining processes, and providing enhanced insights into spending data. “The innovation in fintech has been unbelievable and continues to be that way,” says Alexis. “There’s more to come.”

Curious about how AWS can help kickstart your Fintech startup? Join our latest Global Fintech CTO Fellowship cohort launching April 2023!

Scaling rapidly with AWS—How SEON achieved 3x growth for 3 years running

Bilal Dayeh — Mon, 06 Feb 2023 14:40:11 +0000

Scaling a startup successfully involves increasing profit margins exponentially while keeping costs low. Most startups combine a variety of approaches to scale, based on their growth stage and needs. Techniques to scale include finding processes that work and applying them across the board, focusing on customers and building a product that is in high demand, and harnessing AWS cloud technology to move fast and optimize your costs.

SEON, a Hungarian fraud prevention startup founded by Tamás Kádár and Bence Jendruszák in 2017, is a model of successful startup scaling: Without major refactors of their architecture, SEON has scaled rapidly for three consecutive years, achieving triple growth each year by building on cloud services offered by AWS. In 2021 alone, SEON more than tripled its annual recurring revenue, grew its headcount by 4X, and opened new offices in Austin, Texas and Jakarta, Indonesia.

AWS has been a key technology partner in enabling SEON’s exponential growth.

Building a scalable and cost-optimized architecture on AWS

Chief Architect Adam Berkecz, SEON

A key driver of SEON’s successful scaling, according to their Chief Architect Adam Berkecz, is their use of over 30 AWS solutions regularly.

“The traditional approach of provisioning environments without AWS cloud solutions is expensive and has the hidden cost of time needed to launch. With AWS, we have more than 100 engineers delivering customer value on a diverse technical portfolio,” explains Adam.

The stars of SEON’s architecture include AWS solutions such as Amazon Elastic Compute Cloud (Amazon EC2), Amazon Relational Database Service (Amazon RDS), Amazon API Gateway, and AWS Lambda, which allow them to handle real-time transactions for more than 5,000 customers.

A simplified view of SEON’s AWS architecture for a single region.

The flexible scaling of these AWS solutions enables SEON’s architecture to thrive, even during elongated periods of high load. This flexibility was on display when SEON launched the fraud browser detection feature in their device fingerprinting solution and enabled it instantly for their customers’ millions of end-users. SEON served over 10,000 requests in the first minute without any scalability issues.

In addition to granting flexibility, SEON’s AWS solutions help them to keep costs predictable. By employing AWS Savings Plans and Amazon EC2 reserved instances, SEON ensures that they are not overpaying for their compute resources. In addition to that, SEON stays on top of their spending by regularly monitoring AWS Cost Explorer and its granular view on linked accounts, services, and usage types. Finally, for infrequent and event-driven compute tasks, SEON opted to go serverless by using AWS Lambda: This allows them to save more on costs and at the same time not need to provision any instances, nor manage them.

Key tips for enabling rapid growth with AWS

1. Keep it simple. When looking for a minimum viable product (MVP) or a market fit with a new product offering, stick to the most easy-to-use AWS services like AWS Elastic Beanstalk. Simple yet powerful offerings like Elastic Beanstalk enable your organization to focus on building products rather than invest time in managing services. For SEON, it is important that developers stay as productive as possible to propel the company’s growth.

“With AWS Elastic Beanstalk and Lambda solutions, we are able to have developers working in various languages (Java, TypeScript, Python, Golang, and others) while focusing on writing code and not on managing servers and databases. With this approach, we can spin up new environments in minutes,” says Adam.

2. Invest in a multi-AZ and multi-region architecture. When clients send SEON’s tools a transaction to review, a customer at the other end is hoping to sign up for a new service or place an order. Every second that passes will affect their overall customer experience.

By investing in multi-AZ and multi-region architecture, SEON is able to maintain approximately 2–3 second response times around the world. Furthermore, SEON maintains excellent service availability even in the rare cases of service degradation in one zone or another.

SEON’s multi-region AWS architecture

3. Support experimenting with new services. SEON’s architecture is constantly evolving. This evolution is possible as their leadership supports innovation and testing new AWS technologies. By using a sandbox account, SEON’s engineers can build small architectures and proof-of-concepts that may be eventually propagated into production. For example, experimenting with serverless technologies like Lambda and different flavors of RDS databases allowed SEON to realize that they can improve their application architecture with these changes and they consequently mirrored them in their production environment.

SEON’s response times dropped from roughly 1500 milliseconds to around 600 milliseconds after incorporating native AWS services like Amazon Simple Queue Service (Amazon SQS) into their architecture.

What’s next for SEON?

Having raised $94 million in Series B funding in April 2022, SEON is looking to expand its presence in North America, Latin America, and the Asia Pacific region. SEON continues to build partnerships with leading ecommerce platforms, heighten product functionality, and integrate additional data sources to help customers better fight fraud.

“With AWS continuously providing and updating futuristic services for AI, containerization, and message streaming, we do not see ourselves slowing down,” says Adam. “Managed services like Amazon Aurora and managed Kafka are on our technological roadmap, and we look forward to what we can accomplish further with them.”

How Amazon SageMaker helps Widebot provide Arabic sentiment analysis

Mohamed Mostafa — Fri, 03 Feb 2023 14:28:03 +0000

Startups are familiar with the importance of creating great customer experiences. Sentiment analysis is one tool that helps with this. It categorizes data as positive, negative, or neutral based on machine learning techniques such as text analysis and natural language processing (NLP). Companies use sentiment analysis to measure the satisfaction of clients for a target product or service.

Sentiment analysis can be particularly challenging to accomplish on Arabic end-users: People across the Middle East and North Africa (MENA) region speak more than 20 dialects of the Arabic language, with Modern Standard Arabic being the most common language.

In this blog post, we explain how Widebot uses Amazon Sagemaker to successfully implement a sentiment classifier. Widebot is one of the leading Arabic-focused conversational artificial intelligence (AI) chatbot platforms in the MENA region. Their sentiment classifier supports Modern Standard Arabic, as well as Egyptian dialect Arabic, with high accuracy when tested on multiple datasets from different domains.

Widebot’s model can be easily tuned after being given a few hundreds of samples from the new domain or dataset. That makes the solution generic and adaptable to different domains and use cases.

The characteristics of a successful chatbot

Chatbots are a useful tool for managing and improving customer experiences, as well as automating tasks so that employees can focus on work critical to their company’s mission. Startups, in particular, are familiar with the value of using managed services so that they can spend their time on the tasks that matter most to their success.

It’s important for chatbots to quantify satisfied or unsatisfied customers, as well as document the conversion rate from satisfied-to-unsatisfied (or vice versa). To meet these requirements, Widebot’s solution:

Helps users to analyze the performance of their chatbot system
Improves the decision-making of the chatbot
Aids other downstream machine learning (ML) models

Widebot chatbots use sentiment analysis to improve customer interactions.

Technical challenges of building sentiment analysis

Widebot data scientists are always innovating to enhance and optimize their deep learning models to keep up with their customers’ growing expectations. To better serve their Arabic chatbot customers, they worked to develop a new solution for Arabic sentiment analysis deep learning models.

The challenges of this included:

Model scalability
Response time
Massive concurrency requests
The running cost

As is the case for many startups, they initially deployed the model on self-managed infrastructure and general-purpose servers. However, as their startup grew, they couldn’t efficiently scale the model to accommodate for the growing data and spikes on concurrent requests.

Widebot began looking for a solution to help them focus on building the models quickly, without devoting undue time to managing and scaling the underlying infrastructure and machine learning operations (MLOps) workflows.

Model deployment on Amazon SageMaker

Widebot chose SageMaker because it provides a broad selection of ML infrastructure and model deployment options to meet all their ML inference needs. SageMaker makes it easy for startups to deploy ML models at the best price performance.

“Fortunately, we found that Amazon SageMaker gives us full ownership and control throughout the model development lifecycle. SageMaker’s simple and powerful tools allow us to automate and standardize the MLOps practice to build, train, deploy, and manage models more easily and quickly than was possible through our self-managed infrastructure,” said Mohamed Mostafa, co-founder and Chief Technology Officer (CTO), Widebot.

The Widebot team are now able to focus on building and enhancing their ML models to meet their customer expectations, while SageMaker takes care of setting up and managing instances, software version compatibilities, and patching versions. SageMaker also provides built-in metrics and logs for endpoints to keep monitoring the model health and performance.

Amazon SageMaker Inference Recommender helped Widebot to choose the best compute instance and configuration to deploy their ML models for optimal inference performance and cost. SageMaker Inference Recommender automatically selects the compute instance type, instance count, container parameters, and model optimizations for inference to maximize performance and minimize cost.

Widebot also uses various AWS services to build their architecture, including Amazon Simple Storage Service (Amazon S3), AWS Lambda, Amazon API Gateway, and Amazon Elastic Container Registry (ECR):

Widebot architecture diagram

Widebot was looking for a solution to securely publish the ML models they developed for their customers as an API endpoint. They used API Gateway, a fully managed service, to publish, maintain, monitor, and secure the API endpoints of the ML models deployed on SageMaker. API Gateway is used as an external-facing, single point of entry for SageMaker endpoints that makes them easily and securely accessible from clients.

Clients interact with the SageMaker inference endpoint by sending an API request to the API Gateway endpoint. The API Gateway maps client requests to the corresponding SageMaker inference endpoint and invokes the endpoint to obtain an inference from the model. Subsequently, the API Gateway receives the response from the SageMaker endpoint and maps it back, in a response sent to the client.

Solution overview

How did Widebot build a successful new solution for Arabic sentiment analysis deep learning models? Here are the steps they followed:

Datasets collection and preparation

Collect tens of thousands of data samples from different data sources (both public and in-house).
Review the datasets carefully, apply data labeling, and improved the data quality by removing irrelevant samples.
The data team conducts an annotation process, using Amazon SageMaker Ground Truth to annotate enough samples from different domains and writing styles to enrich the dataset used.
Send samples through the preprocessing pipeline, before training the model using deep learning to classify the input text as positive, negative, or neutral, with the probability of each.

Building and training the model

Use a Convolutional Neural Network (CNN) model trained using Keras and TensorFlow.
Apply many iterations to test different preprocessing pipelines, architectures, and tokenizers, until reaching the architecture that yields the best results on different sample datasets and from different domains.
Use a native preprocessing pipeline developed in-house to remove unnecessary information from the text: dates, URLs, mentions, email addresses, punctuation (except for ‘!?’), and numbers.
Apply Arabic text normalization steps, like stripping diacritics and normalizing some letters that users used interchangeably, like (ء أ ئ ؤ إ) or yaa (ي ى) or other characters.
Apply light stemming on the text that removes some suffixes and prefixes and reduces some inflated words into their stem (for example, (التعيينات) reduced into (تعيين)).
Save the model, preprocessor, hyper-parameters, and tokenizers using serialization and export them as .h5 and .pickle files.

Deploying the model on Amazon SageMaker

Wrap the model into an API, the prediction endpoint. That endpoint accepts JSON input from the end user and transforms data into an easier data structure, cleans it, and returns the sentiment results of the input data.
Create a Docker image that contains the code, all dependencies, and instructions required to build and run the components in any environment.
Upload the model artifacts to an Amazon S3 bucket and the Docker image to Amazon ECR.
Deploy the model using SageMaker, selecting the image location in Amazon ECR and the artifacts URI in the Amazon S3 bucket.
Create an endpoint using SageMaker and leverage API Gateway to publish the endpoint to their clients.

Type and volume of data

To build their model, Widebot’s data consists of approximately 100,000 different messages for training and 20,000 messages for validation and testing. The messages:

Came from different industries, such as e-commerce, food and beverage, and financial services.
Included reviews for different services or products. For example, hotel reviews, booking reviews, restaurant reviews, and company reviews.
Ranged in tone from very formal language to the use of severe profane words.
Were written in both Egyptian dialect and Modern Standard Arabic.
Were classified into one of three classes: negative, neutral, or positive.

The following table shows sample messages:

Example	Sentiment	Confidence
الخدمة لديكم مناسبة “Your service is good”	positive	0.8471
شكرا لحسن تعاونكم “Thank you for your cooperation ”	positive	0.9688
الخدمة والتعامل لديكم دون المستوى “Your service is substandard”	negative	0.8982
حالة الجو سيئة جدا “The weather is very bad”	negative	0.9737
سأعاود الإتصال بكم وقت لاحق “I will contact you later”	neutral	0.8255
أريد الإستعلام عن الخدمات “I want to inquire about the services”	neutral	0.9728

Results Summary

Widebot tested their model against different Arabic text datasets in various dialects. These metrics were measured using datasets with thousands of samples. The F1-score is used to measure the model’s accuracy with the different datasets. The macro and weighted averages of the F1 score are used to measure overall precision and performance.

The model accuracy

The testing dataset (20,679 samples in the ratio 5004:1783:13892)

Negative F1	Neutral F1	Positive F1	Overall accuracy	Macro average	Weighted average
89.9	79.4	95.1	92.5	88.1	92.5

The model response time

Widebot measured the response time using the average (AVG), minimum (MIN), and maximum (MAX) seconds per response (sec./response):

AVG: 0.106 sec./response
MIN: 0.088 sec./response
MAX: 0.957 sec./response

The following compares the response-time metric between using a general-purpose compute platform and using Amazon SageMaker for model hosting, when deploying the same datasets with an average payload size of 2 KB.

Total response time	General compute platform (EC2 instances: p2.xlarge)	Amazon SageMaker (SageMaker instances: ml.m4.xlarge)
Average	0.202 sec./response	0.106 sec./response
Minimum	0.097 sec./response	0.088 sec./response
Maximum	8.458 sec./response	0.957 sec./response

The model concurrency

The model was able to handle 1,000 concurrent requests served on average in 164 milliseconds.

The model concurrency calculated metrics. These metrics measure the throughput and efficiency when handling concurrent requests.

Conclusion

This post shows how AWS services helped Widebot to build a comprehensive solution to extract sentiments from chat text in different Arabic dialects, using a deep learning model hosted on SageMaker.

SageMaker helped Widebot innovate faster and deploy their sentiment classifier to solve the complex ML problem of extracting sentiments from Arabic conversational text and to publish this as a public RESTful endpoint for clients to access easily and securely via the API Gateway.

This approach could be useful for many similar use cases, where customers want to build, train and deploy their ML model on SageMaker and then publish the model inference endpoint for their customers in a simple yet secure way, using the API Gateway.

If you are interested in reading more on linguistic diversity and how to fine tune pre-trained transformer-based language models on Amazon SageMaker, you can read this blogpost.

CareCoPilot founder Alyse Dunn wins big after AWS Impact Accelerator

Chalaire Miller — Wed, 01 Feb 2023 18:03:01 +0000

To celebrate Black History Month, AWS Startups is featuring posts throughout February highlighting the contributions of Black builders and leaders in tech. Above all, these individuals inspire, empower, and encourage others—especially those historically underrepresented in tech—to prove what’s possible.

Just before Thanksgiving in 2014, CareCoPilot founder Alyse Dunn knew she had to make a change. She and her sister had spent the past two years managing their father’s care as he dealt with multiple sclerosis. Then, shortly before his death, their mother received an Alzheimer’s diagnosis.

“I knew I needed a job that could provide me with more money and more flexibility to take care of her,” Alyse recently told AWS Startups. “I didn’t have the resources that I needed to take care of my dad in the way that I wanted, so I made this drastic career change to avoid that mistake with my mom.”

Coding for caregiving

Without any prior technical experience, Alyse taught herself to code and pivoted to a career in software engineering. While it did provide her with the financial resources and flexibility that caregiving necessitates, she still describes the process of caring for her mother as “crushing.”

“In the course of those eight years,” she says, “We never found anything that made our experience caregiving even 10% better.”

Alyse vowed to change that. Like many founders, her investment in her idea was both professional and personal. She recognized a critical hole in the market and knew that the growing elder care crisis would make her product all the more essential. But she was also determined to follow in the footsteps of her parents—both were physicians—and dedicate herself to designing a product that could help and heal others.

Thanks to both her software engineering experience and the skill set she had developed while caregiving, she created CareCoPilot. The web and mobile app makes it easier for caregivers to discover and access the resources that can save them time and money.

Additionally, the app allows caregivers to join a rewards program that they can eventually use to find, book, and finance many of the resources needed to manage care, like medical help, legal expertise, and home healthcare items.

The right call at the right time

Alyse funded her early days by quitting her job at Venmo and cashing out her stock options. Just two hours after quitting, she got a call that she’d been accepted to her first accelerator. It came with a $25,000 check, but in the weeks that followed the program, she still wasn’t raising as much as she’d hoped.

She was staring down a precarious financial situation when lightning struck the second time—a call from Denise Quashie at AWS Startups, inviting her to join the inaugural AWS Impact Accelerator for Black Founders.

“It was really and truly an answer to my prayers,” says Alyse. “I would literally not be here today, helping families go through this, if AWS had not stepped in.”

Taking advantage of the AWS Impact Accelerator Program

Alyse went into the program with a few different goals. She had been an early adopter of AWS, so she was skeptical about how much she’d be able to accelerate her product on the technical side. But she was pleasantly surprised to learn about tooling services she didn’t know existed that are now boosting her productivity and protecting her infrastructure.

The program also exceeded her expectations when it came to the camaraderie and support she experienced. All 25 participants were Black American founders, and that kinship allowed Alyse to open up, accept and give support, and be more authentic than she felt able to be in other accelerator programs.

Additionally, she was able to work with AWS mentors to hone her pitch. The hard work paid off weeks after the AWS program, when Alyse took a leap of faith by flying to Atlanta for a pitch competition hosted by the Fearless Fund.

Guided by the advice of AWS mentors who encouraged her to drive engagement with funders by putting them in her shoes, she gave the real, raw version of her startup journey. Her emotional pitch included difficult admissions that would resonate with anyone who has managed care, as well as a plea for those who haven’t experienced it to determine where they would find the time and money to take on the role.

Landing a winning pitch

The pitch landed—Alyse went home with the second place prize and a $325,000 investment in CareCoPilot. With the cash boost, she is focusing on increasing signups, looking ahead to new pitch competitions, and gathering feedback to improve CareCoPilot.

“I think that one of the mistakes that a lot of founders make is that they think that the winner of the startup game is the founder with the best idea out of the box,” says Dunn. “That’s not really the case. The winner of the startup game really is, honestly, whoever can get close, be really open to feedback, be able to iterate cheaply and quickly and not run out of money before you get it right.”

She understands her startup path may not be seamless. But on her desk, she has a picture of her parents she can turn to when she needs motivation to keep going. “They are my inspiration because they were extraordinary parents,” says AlyseBA. “Their experience, me and my sister’s experience, it wasn’t for nothing. It was for something, and it was for this.”

Explore more content that celebrates the achievements of Black innovators, such as:

AWS Startups Twitter stories of inspiring, innovative, and game-changing Black founders
How Resilia is helping nonprofits to build capacity through SaaS
Shining a light on Black excellence: three founders share their stories
How Eyegage scaled their life-saving app via Impact Accelerator partnership

What Amazon CTO Werner Vogels’ predictions for 2023 mean for startups

Chris Munns — Thu, 26 Jan 2023 20:28:26 +0000

Last year, we published our first ever startup-specific take on Amazon CTO Werner Vogel’s yearly predictions for technology. As Amazon’s CTO, Werner has a wide view across industries and countries, seeing both what’s being invented by hot startups, as well as seeing what’s happening in the world’s largest enterprises. Based on this, he publishes predictions for the year ahead and beyond.

For 2023, Werner’s 5 predictions cover topics both specific to certain industries and technologies, but also those that impact the lives of many around the world. These predictions highlight:

How human actions and behaviors can be influenced by artificial intelligence (AI) and analytics
Simulated research in virtual worlds is impacting the real world
Energy innovation is helping us tackle growing energy demand
Supply chains are being reinvented
Custom silicon and hardware in the cloud enables these predictions to come to life, and more

Join me again in diving into these predictions, and see how I think they’ll help startups prove what’s possible in 2023 and beyond.

1. Cloud technologies will redefine sports as we know them

Werner’s first prediction discusses how technology is gaining yards (pun intended) in sports. Since the early 2000s, statistics and data models have shifted how teams are built and how athletes play.

Technology like wearables and AI models can assess athletes’ performance and and track a team’s plays across an entire field. With the data they derive, coaches can determine how and when certain plays perform best, or fans can see real-time stats while watching a game.

For startups, I think this highlights how machine learning, AI, Internet of Things (IoT) devices, and analytics can help humans improve almost any task. Being able to track the actions and behaviors of top performers and finding room for minor improvements can add up to big business benefits, which would be just as valuable as an AI model that shows what plays have better chances of success.

I also see a strong overlap with Werner’s next prediction of how simulated worlds can impact the real one–how could we could simulate data of real human performance to suggest improvements?

2. Simulated worlds will reinvent the way we experiment

We often think of virtual worlds and virtual reality (VR) as a way to escape the real world. Putting on a VR headset and playing an immersive game can be a fun and exciting new way to experience virtual worlds. But there can be another very real and valuable aspect to these virtual worlds: virtual simulations.

Today, virtual simulations are used to improve race cars, predict weather, model the stock market, and more. As Werner writes in this second prediction, this is just the start of what’s possible as new technologies make simulations almost as valuable as testing in the real physical world. By combining real-world data, models built against that data, and then simulation technologies such as AWS SimSpace Weaver, companies are able to build realistic test scenarios that would closely match what costly and complex physical testing could result in.

For startups, the opportunity to build and invent in this space is virtually unlimited. Already, brands use augmented reality (AR) tools to show what a piece of furniture might look like in your home. But what if you could predict how light and sound might carry in a room around it? Or, how might the layout of a store impact product discovery and sales? What if landscaping design’s impact on irrigation and water flow could lead to more fruitful gardens and plantings? What if a middle schooler could model the impacts of a derby car design to win their school’s Soap Box Derby? In a virtual world, the constraints of physics and natural elements can be modeled and tested to drive impact.

3. A surge of innovation in smart energy

With energy demand at historic highs in many places around the world, the need for better and smarter energy systems is clear. Werner’s third prediction is that of rapid development in this space.

Growth of renewable-powered energy sources and batteries power everything from trucks to homes. These present opportunities for startups looking to innovate at every stage in the generation, transport, and consumption of energy. Other possibilities include finding ways to better manage the overall power consumption of in-home appliances, both in alignment with peak demand cycles, but also by reducing functionality to optimize power usage.

The opportunity for startups to power innovation in this space and bring to bear the benefits of machine learning, AI, and IoT, are significant. What if personal habits could be tweaked in ways to enable better energy consumption or if devices could tell you when they need to be charged? What if a small business could benefit from smarter in-store appliances and devices that knew when the shop closed and when it would open for business? And what if the modeling of battery capacity and power could be used to shape the capabilities of devices such that they need less charging than a competitor’s?

In the space of energy innovation, these examples show that there are many opportunities for startups to charge ahead.

4. The upcoming supply chain transformation

The global COVID-19 pandemic that started in 2020 got us all thinking how some of our goods are manufactured, shipped, and purchased across the globe. Werner’s fourth prediction speaks to how the modernization across several fronts in supply chains will bring an improvement to day-to-day life. From route optimization for cruise ships, to the advances in autonomous trucking, to how robots can help better organize and select products off shelves, the possibilities for innovation will be huge.

For startups, the opportunities in supply chain automation and invention seem like an open road. From large to small stores, to local or transcontinental shipping, startups can make an impact through IoT sensors, machine learning, AI, or data analytics.

And startups don’t need to just focus on boats, trains, trucks, and planes. The innovation possible in last-mile supply chain capabilities remains a very active area. With many companies investing in how to-door delivery is done, to businesses who need flexible B2B options for on-demand delivery capacity, this remains an unsolved facet of the supply chain, per Werner’s prediction.

5. Custom silicon goes mainstream

As Werner’s fifth prediction highlights, the days of the generic processor being “good enough” for most workloads are over. Custom silicon and specialized hardware, whether they’re in the cloud processing machine learning data, or at the edge in a device processing signal data in near real-time, is now the norm. Improvements in low-powered CPUs that can handle most generalized workloads are also helping customers save money and shrink instance fleet sizes.

For example, AWS’s Graviton3-based instances use less energy for the same performance as comparable Amazon Elastic Compute Cloud (Amazon EC2) instances with non-Graviton CPUs. Similarly, the purpose-built nature of processors designed for AWS Tranium and AWS Inferentia instances improve the cost-to-train and inference performance for machine learning workloads.

For startups, few will ever need to design their own silicon or build custom hardware, but they will reap the benefits provided by using these tools. By aligning the best compute resources for generic workloads, such as databases, web applications, or media transcoding, startups will find the “goldilocks” of right-sizing: not too big, not too small, but just right for their workloads at any point, with the appropriate cost and performance required.

Faster and lower cost machine learning offered by cloud technology will allow startups to experiment more often, fine tune their models even better, and find ways to reduce their costs while doing so.

Conclusion

As these technology predictions highlight, we continue to see broad advancements in AI, machine learning, virtual environments, and hardware mixing to enable exciting new business ideas.

Technology also continues to enable a more equitable world. Smaller and lower powered personal devices that enable connectivity and communication to the internet are in the hands of billions.

More than ever before, startups have access to the tools needed to build the next great thing. We look forward to seeing what you build in 2023!

Entrepreneurship shines in the 2022 AWS University Startup Competition

Qxhna Titcomb — Thu, 26 Jan 2023 18:45:21 +0000

It’s no secret that some of the most successful startups were founded by members of the university community: from Ava Labs to Anyscale and InsightFinder, to name a few. At Amazon Web Services (AWS), we believe this is because students and faculty are often creative thinkers who are willing to take risks and collaborate with their peers—all essential qualities in a founder.

With this in mind, AWS Startups launched the University Startup Competition to find and support student and faculty entrepreneurs as they build and launch their ventures. The competition is in partnership with Amazon Launchpad, a program that supports entrepreneurs by providing resources, expertise, and global support to help showcase and deliver unique products to Amazon customers.

Applicants to the University Startup Competition are associated with a US-based university as an undergraduate or graduate student, faculty member, or staff member.

Now in its third year, the 2022 AWS University Startup Competition received over 1,000 applications between September and November from startups across 300+ university campuses.

Applications per state.

Prizes and categories for the AWS University Startup Competition

To encourage startups to keep testing and building, all applicants qualify for up to $1,000 in AWS Activate Portfolio credits. The thirteen finalists qualify for up to $5,000 in AWS Activate Portfolio credits, additive to any of the category prizes:

1st place – $20,000 cash + qualify for up to $100,000 in AWS credits*
2nd place – $10,000 cash + qualify for up to $100,000 in AWS credits*
3rd place – $5,000 cash + qualify for up to $100,000 in AWS credits*
Top physical consumer products startup – $10,000 cash + up to $10,000 in marketing support, both from Amazon Launchpad
Top AI/ML startup – qualify for $10,000 in AWS credits*
Top HCLS startup – qualify for $10,000 in AWS credits*
Top web3 startup – qualify for $10,000 in AWS credits*
Top fintech startup – qualify for $10,000 in AWS credits*

*Pending AWS Activate eligibility requirements here.

The competition process

The AWS University Startup Competition consists of three rounds of evaluation. For each round, AWS Startups hand-picks a team of evaluators based on their experience as founders, investors, and operators.

The stages of the competition.

Selecting the semi-finalists

To narrow down the 1000+ applications to 150 semi-finalists, in December the first group of evaluators reviewed each application with the following criteria in mind:

Idea
Product/Service
Preparation for launch
Execution plan
Progress/Traction

A second group of evaluators selected 13 of the semi-finalists to participate in the final round: a virtual live-stream pitch event. In addition to the original criteria, the finalists were chosen for their ability to articulate that their startup offers a “need to have” solution instead of a solution that is “nice to have.”

Live-streaming the finalists’ pitches

On January 19th, 2023, the AWS Startups University Team streamed the final pitch event via a Zoom webinar. Each finalist presented virtually to a panel of four judges.

Judges:

Caroline Toch, Operating Principal, Dorm Room Fund

Jin Kim, Senior Vice President, Alumni Ventures

Erik Pavelka, Senior Manager – Business Development; Early Stage Startups, AWS

Senior Product Marketing Manager; Amazon Launchpad, Amazon" width="176" height="202">

Stephanie Danner, Senior Product Marketing Manager; Amazon Launchpad, Amazon

Startups were judged on clarity of their solutions, progress in their journey, and vision to scale their business.

The judges shared the most impressive things they saw at the final pitch competition were strong founder market fit, articulate explanations of how startup solutions connected to problems, and confident and prepared speakers. In particular, the diversity among applicants and among the schools represented was a highlight of the competition.

For applicants who plan to compete in next year’s University Startup Competition, Stephanie Danner advises, “Obsess over your target customer. Gather insights, data, and anecdotes to ensure you deeply understand their wants and needs, and design around them. Be mindful of what your competitors are doing, but don’t let them distract you from your ultimate end goal!”

Meet the winners

Without further ado, meet the overall competition winners and the category winners. These are some of the best and brightest startups within the US university startup ecosystem.

First place (and also the category winner for “Top healthcare / life sciences (HCLS) startup”)

Hubly Surgical, from John Hopkins University, built the Hubly Cranial Drill to provide safer medical drilling during neurosurgery. Unique features on the single-use and battery-powered drill include auto-stop, force indication, and visual feedback. The drill improves surgical safety while reducing patient complications and operating room dependence.

“I was so honored to compete alongside all such impressive student founders and absolutely astounded to have won! The AWS program officers have been (and continue to be) incredibly supportive throughout the entire process. The amount of hard work they put into making this competition successful blew me away. They went above and beyond to ensure that each team had the resources and support it needed to succeed. And to the expert judges who took the time to come out and review our pitches: thank you for lending your expertise and knowledge. I am extremely grateful for the opportunity to learn from your experience. Thank you again for this honor! I am excited to put this funding toward our early clinical pilots—a crucial step in modernizing neurosurgery for the better.”—Casey Grage

Co-founders:

The startup will use their prize to fund their first in-patient pilots in Chile. This includes building and shipping the Hubly drills, as well as sending a team member to Chile to organize and train the participating neurosurgeons.

Second place (and also the category winner for “Top fintech startup”)

Sotira, from UC Berkeley, is a customizable tool that integrates with top e-commerce marketplaces to provide sellers and resellers up-to-date analytics and insights. Sotira helps sellers and resellers to be more profitable and to streamline their sales, profits, and pricing.

“One key takeaway and learning is the value of starting to build in school and all the resources to build, test and iterate that students have access to. ”—Amrita Bhasin

Co-founders:

The Sotira team will put their prize toward the $1.5 million round they are currently raising. They plan to allocate 70% of the funds to hiring and growth, 15% to marketing, 10% to security and storage, and 5% to legal.

Third place

Boston Quantum, which hails from the Massachusetts Institute of Technology (MIT), is working to disrupt the financial industry with their enterprise quantum computing software. Their Vision software locates arbitrage opportunities in systems with over 10 currencies, offers interface via easy-to-integrate APIs, and obtains results that allow users to trade in milliseconds.

“This year’s finalist cohort had both an incredible range of focus and depth of expertise. Clearly, students across fields want to get their hands dirty and build ventures, while still in school. This experiential learning even seems to be an essential part of the education of an aspiring entrepreneur. With the right resources and support networks in place, universities can nurture and kickstart generations of entrepreneurs. I’m excited to see how today’s student entrepreneurs will shape tomorrow. ”—Shantanu Jha

Co-founders:

The Boston Quantum team will use their prize to continue bootstrapping, quantifying their value proposition, and onboarding paying customers. Looking forward, they plan to hyper scale.

Top physical consumer products startup

Ceres Plant Protein Cereal, from Tulane University, has created a vegan, keto, non-GMO, and gluten-free cereal that is actively good for the planet. Each serving of cereal contains 20 grams of sustainably sourced plant protein and zero sugar or artificial sweeteners. They ship sustainably and have a carbon footprint lower than most animal-based breakfasts.

“Through the AWS University Startup Competition, I got a sneak peek of tomorrow’s revolutionary companies. It’s empowering to share the stage alongside many companies working to tackle important social and environmental problems through innovative concepts.”—Rich Simmerman

Co-founders:

Ceres will use their prize to support their ingredient-sourcing and production process, lower their cost of goods sold, expand their business on Amazon and Thrive Market, and market their cereal.

Top artificial intelligence / machine learning (AI/ML) startup

S-3 Research, from the University of California, San Diego, is a research-as-a-service company that meets public health challenges with custom technology and human research. They have worked to prevent illegal online sales of opioid and fentanyl, helped fight the COVID-19 pandemic, and improved health equity with their S-3 engine, which combines data generation with analysis and visualization.

“The major takeaways and learning we got out of the AWS University Startup competition is that there is a huge diversity of solutions that can be built on or leverage AWS architecture, and that flexibility leads to an accessible environment for innovation that in the past would have been much more resource intensive.”—Tim K. Mackey

Co-founders:

S-3 Research plans to use the prize to integrate their technology stack into their AWS infrastructure and tackle underserved societal challenges.

Top web3 startup

CryptoClear, from the Massachusetts Institute of Technology (MIT), is a software-as-a-service company that provides crypto ecosystem ratings and portfolio analysis for investors and institutions. Their platform uses big data and AI to help people quickly identify relative value and manage risks.

“The AWS competition is an incredible opportunity for us to showcase our startup in front of a national audience. The tremendous support that we received from AWS Startups throughout our startup journey and the final recognition go above and beyond what we can express in words. ”— Anne Wang

Co-founders:

The CryptoClear team will put their prize toward a $1 million seed round, which will support their product research and development, IT deployment, and data sourcing. The AWS credit will help with their data storage and cloud computing.

Looking ahead to the 2023 University Startup Competition

Congratulations to all of the winners for working hard, showing up, and disrupting industries in the pursuit of proving what’s possible.

Do you want to participate in next year’s AWS University Startup Competition? Get started today—wherever you are in your startup journey, AWS has you covered:

AWS Activate: Find tools, resources, content and expert support to accelerate your startup at every stage.
AWS Startup Lofts: Get founder assistance, from inception to IPO.
AWS Cost Explorer: Visualize, understand, and manage your AWS costs and usage over time.
AWS Startup Programs: Discover programs that connect your startup to new opportunities, from events and accelerators to technical partners and more.

Evolutionary architectures series, part 1

Zoran Nakev — Mon, 23 Jan 2023 17:29:06 +0000

A team evaluates an initial startup idea.

“I’ve got this great idea!”

Every startup begins as an idea. Before you start worrying about funding or staff or distribution or any of the other myriad things, you have your fresh, new idea—a product or service that you think has potential.

If your idea will rely on the cloud, you’ll need a cloud architecture. This blueprint will help usher your great idea into reality and, if built well, can evolve alongside your business as it grows.

To help you build a robust blueprint for your idea, this four-part series, Evolutionary Architectures, will show you how one company, the aptly named Example Startup, puts their idea into practice. In part 1, we’ll see how they built a minimum viable product (MVP) to test customer interest and market fit. Later in the series, we’ll look at how their designs and decisions evolve as they move through startups lifecycle to deliver a fully-fledged, scalable, secure, highly available, and redundant solution.

Delivering your first MVP

The first few product deliveries by a startup usually follow a phased approach. They’re dictated by funding, time, resources, team size, and knowledge and experience.

In this stage it’s extremely important to not let perfect get in the way of the good and to deliver simple but functional solutions. To do this, you’ll need to know how to identify 1-way door and 2-way door decisions, fail fast and pivot when necessary, manage cost, and speed up time to market.

Let’s check in on Example Startup and see how they approach this process.

The idea

Example Startup’s idea is to create a “fantasy stock market.” They used fantasy sports leagues as a baseline and applied the idea of stock market investing. They envision holding four “tournaments” over the course of a year.

At the beginning of every quarter, a new cohort of investors starts with the same amount of funds. The fantasy stock market allows these investors to make their investment choices (based on companies and symbols from real stock markets) over the following 3 months. At the end of the quarter, participants are ranked and winners are announced.

The preparation

To make their idea a reality, the two founders bootstrapped their startup: they gathered their savings and borrowed money from friends and family.

One founder, an experienced developer, got 3 months leave from her job. This allows her to focus on the technical solution, which is helpful. However, it also defines the timeline for their first delivery.

Now, in just three months, she and her co-founder, who has a background in finance, must decide which features to include in the MVP and build their product. To get started, they decide on 1) what features are absolutely necessary for a usable product and 2) what features will allow them to measure market fit and customer interest.

They decide on the following:

Stock market analysis for the product

An import process for real-world stock market symbols/companies that investors can trade on
Daily market prices feed
Signup mechanism for users
Portfolio management user interface (UI)
Daily process for end-of-day portfolio calculations
A daily process that calculates rankings

The build

After defining their scope, it’s time to make some technical decisions about which technologies and components the fantasy stock market needs. Then, they’ll create an implementation plan with milestones for the MVP launch.

Framework

As a developer, one of Example Startup’s founders has experience in React, a JavaScript library for building user interfaces. Considering that a big portion of the MVP deliveries involve UI development, she thinks AWS Amplify looks like a great fit. With Amplify, the team gets built-in support for building and hosting React.js applications with lots of reusable components. Amplify can help with the backend as well—it can manage different databases like Amazon DynamoDB, a great flexible option to start with, and it can use AWS AppSync to easily connect the front-end with data sources and develop the business logic.

Domain

With the framework taken care of, it’s time to get a domain name. Amazon Route 53 helps Example Startup to configure a DNS (Domain Name System) service that has good integration with the services they were already using, as well as with the process of domain registration.

Experimentation and Cost Management

The team is able to meet most of their initial needs by simply picking an AWS service that aligns with their use case. The breadth of AWS services allows Example Startup to quickly experiment with multiple options and make decisions based on the experience.

Although many of the AWS services have a free tier, some of Example Startup’s experimentation may be overly enthusiastic. When the first month’s bill arrives, the team realizes that they need to pay more attention to cost. Like in many other cases, there is an AWS solution for that: they start using the free service AWS Budgets. It helps the team improve on their planning and cost management and define alerts that conveniently bring to their attention anything that might not align with their expectations.

AWS Budgets sends budget alerts.

Data

A month in, Example Startup already has a lot of the UI and some related features working with sample data. Next, they’ll need batch processes that will do the heavy lifting with some real data.

After finding data sources to provide the information they need, the team wants to automatically ingest the data. Continuing with JavaScript as their programming language of choice, they want to run it with something that makes the operational aspect as simple as possible.

This leads them to AWS Lambda. The team doesn’t want to worry about operating servers and scaling, so they take a serverless approach, using the Schedule AWS Lambda functions using EventBridge tutorial.

With that, as shown in the initial architecture diagram, they have a design in place for the data-related services they need to run.

Initial architecture diagram

Testing

The team is making great progress. Their architecture is growing, and they feel good about the 3-month deadline.

However, as the number of people testing the solution grows, they notice an issue. Someone on the team asked: “How many people are active users and what’s the average number of transactions per user?” But they couldn’t really come up with a meaningful answer. They agree on “winging it” by temporarily running queries directly on the DynamoDB console and start a “wish list” for the next iteration.

The launch

Example Startup made their deadline and launched the MVP. Before long, the team sees a huge list of registrations. They realize they are onto something, but they need help to improve their product and scale their business.

A friend of a friend mentions AWS Activate, a program that offers startups a number of benefits, including AWS credits, AWS support plan credits, and architecture guidance.

The benefits of AWS Activate.

They apply to AWS Activate to get the help they need for the next phase of their journey.

Conclusion

A lot has happened with Example Startup in just a few months. During the process of delivering the first MVP, many startups face challenges that are similar to those that Example Startup overcame.

We will continue with their journey in upcoming blogs in the Evolutionary Architecture series. Learn how their needs, challenges, and goals change as the company builds and scales.

Meet your AWS account team

Justin Plock — Thu, 19 Jan 2023 14:40:36 +0000

As you begin building a company on AWS, you may be contacted from someone at AWS who’s interested in learning more about what you are building and how we can help. That’s us, your account team.

We’re here to help keep you up-to-date on all that AWS offers, coach you on best technical and business practices (such as the AWS Well-Architected Framework), and make sure you’re not overspending on your AWS services.

Let’s get acquainted with the different people on your AWS account team who you’ll work with in your journey building on AWS. This post will provide clarity on each person’s role on the team and how best to leverage them to achieve your business objectives.

Account managers

Account managers (AMs) are your primary point of contact with AWS for business or technical matters. As your startup grows, your account team will be more involved; at AWS, our larger startups sometimes have multiple AMs supporting them. No matter your startup’s size, however, your AM handles your relationship with AWS.

There are a few career paths within account management, but you’ll start out with an Associate AM.

Associate account manager

Associate AMs inform companies about new releases and events related to AWS. If you’ve ever received an email on upcoming webinars or events, it likely came from an associate AM.

Associate AMs reach out to companies who might not have connected with AWS yet and are interested in learning how AWS can help support their business.

“I really enjoy the role of associate account manager because I can introduce customers to the vast amount of resources that AWS offers. It’s gratifying to see firsthand the opportunities that are created from these resources to help our customers build effectively and grow.”—Sarah Houghton, AWS associate account manager

If you’ve never engaged with AWS directly yet, an Associate AM may be your first contact with us. The more information you are willing to share about what you are building (what AWS services you are using or considering), the Associate AM can get you in touch with the proper teams within AWS.

Account manager

As your company grows, you’ll be assigned an AM. They’ll help you achieve your startup’s business objectives using AWS. AMs work with Associate AMs to make sure you’re getting the information you need when you need it.

Jeff Savio, AWS senior account manager

“As an account manager, I partner with leadership at innovative startups building on AWS. I enjoy aligning on their strategic business goals and collaborating across AWS expert teams to help them achieve those goals and grow their business. It is incredibly rewarding seeing this partnership lead to successful customer outcomes: building disruptive product offerings, expanding to new markets, transforming end-customer experiences, and more.” – Jeff Savio, AWS senior account manager

When to engage an account manager

Account Managers should be your first point of engagement with AWS on non-support related issues. If you have a technical support concern, please follow the standard process of opening a support case (after the case has been opened, feel free to let the AM know of the case number so the account team can monitor its progress). Account Managers can help with billing-related questions, cost optimization opportunities, as well as work directly with solutions architects to help with technical-related questions.

Solutions architects

Account solutions architect

Your AM is paired with an account solutions architect (SA), who helps guide technical solutions for your startup. Solutions architects are a free resource for you. Engage us as often and regularly as you would like. We recommend customers engage SAs before you’ve begun building so we can review your use case and recommend the best architecture and AWS services. SAs work with many customers and have seen best practices on what has worked well. SAs can also engage specialist solutions architects and the service teams depending on the level of depth of the conversation. Solutions architects also gather feedback and submit it to the service team roadmaps as product feature requests (PFRs). Ninety percent of our roadmaps are driven by customer feedback.

Specialist solutions architects

Specialist SAs can help with individual AWS services, such as Amazon SageMaker or Amazon Elastic Kubernetes Service (Amazon EKS), or technology domains like security or machine learning. Sometimes, an SA brings in other specialists or service team members, depending on the use case, such as using a service in a unique way that we haven’t seen before.

An SA can also introduce you to relevant partners from the AWS Partner Network (APN), based on the focus of your company and the timeframe available. AWS Consulting Partners work with customers at all stages of the startup lifecycle, so even if you don’t have developers yet, or maybe you need assistance for a specific one-time project, they can help you continue to make progress.

Ben Gruher, AWS solutions architect

“I love being an SA. I get to be a member of my customers’ teams and work side by side with them to solve interesting technical challenges. It’s extremely rewarding to be a part of a customer’s success and see our work implemented in production. I recently had a customer who was interested in finding areas where they could reduce costs. We did an architecture review and were able to find a solution that lowered their networking costs by over 50%.”

—Ben Gruher, AWS solutions architect

Technical account managers

The technical account manager (TAM) helps startups onboard to AWS and plan and build solutions following operational best practices around resiliency and cost optimization. You’ll be able to work with a TAM if your startup purchases an enterprise support plan or through enterprise on-ramp. TAMs can engage subject matter experts and help with case management. A TAM can also help keep your AWS environment healthy by showing you how to manage your AWS spend, optimize workloads, and manage events. As an example, if you know you’ll be launching a new feature or into a new market that will require scaling up, TAMs can help you engage our AWS Infrastructure Event Management (IEM) team to proactive help you plan for the event.

Sudhan Rameshwaran, AWS technical account manager

“I enjoy being a TAM because I can be a customer advocate. Being a technical point of contact, I get to help with operational excellence by providing architecture guidance, best practice recommendations, and proactive operational reviews. Recently a customer reached out to help them with Cost Optimization. After running a cost optimization workshop, I was able to help them with set of actionable items to reduce cost.”—Sudhan Rameshwaran, AWS technical account manager

Service teams

Service teams are an engineering organization within AWS. They design, develop, and operate the 200+ services AWS offers.

Product managers

Product managers within service teams often engage with customers to provide feedback on feature prioritization or gain insights into how AWS can help them offload some of their administrative or operational workload.

If you need to speak with a service team, your Solutions Architect or Technical Account Manager can help you reach them.

“As a product manager, I ensure that the features we build help customers address their business problems. I enjoy defining product roadmaps based on customer input and working with engineers to see these features delivered to our customers.”—Suresh Sridharan, AWS Cognito senior product manager

Business development

Most of our business development (BD) team members are former founders, investors, or operators who want to help companies grow and scale.

BD manager

Amy Chen, AWS senior BD manager with early-stage startups

A BD manager is experienced in a specialized field, such as fintech, healthcare/life sciences, or artificial intelligence and machine learning (AI/ML). BD Managers come to AWS with deep connections in the startup ecosystem and can share their experience navigating the ups and downs of startup life. They can also help with relevant business introductions, partnership or co-marketing opportunities, or go-to-market strategy resources.

“I love meeting founders and supporting them with resources that are a fit for their stage of growth or for their industry. I’ve helped startups secure speaking opportunities at the AWS Summit and re:Invent, meet internal and external experts for product development and fundraising advice, and connect with go-to-market programs to partner with AWS to reach joint customers”— Amy Chen, AWS senior BD manager with early-stage startups

AWS support

AWS Support helps companies get the most out of AWS products and features. Are you a builder with a “how to” question or a problem to troubleshoot? Do you have a service issue? Our experienced technical support engineers work one-on-one to find solutions, fast.

AWS Support complements our SAs and TAMs to provide support cases 24/7 for problems affecting your company’s workloads. You can choose the right plan, based on your company’s needs, with pay-by-the-month pricing that avoids long-term commitments.

“As a cloud support engineer, I enjoy assisting customers to troubleshoot their complex issues; the varying customer issues provides good learning experience. Interacting with customers to understand their requirements can in turn enhance AWS products.”—Subodh Raut, AWS cloud support engineer

Summary

Building on AWS means you can collaborate with people on various dedicated teams. AWS provides business and technical resources to help companies grow, aiming for optimal solutions at minimal cost. We want what is best for the companies building on AWS, even if it means working with products and services originating outside of AWS. We want to earn your trust as long-term partners, not short-term sellers.

Hidden gems from the AWS Startups Blog

Bonnie McClure — Fri, 13 Jan 2023 18:47:05 +0000

It’s the new year – for some people, it’s the ideal time to renew focus on your business. To get you reinvigorated and inspired in the new year, we’ve picked some of our favorite posts from 2022.

Startup stories

How LabVoice + AWS are expanding accessibility in research labs

As a PhD student in biophysics at Yale, Sara Siwiecki spends a lot of time in the lab working on the small details, like peering through a microscope or managing inventories of the chemicals she needs for experiments. But she often encounters obstacles due to outdated lab and equipment designs that don’t accommodate her visual impairment. At Brown, PhD student Gabriel Monteiro da Silva has spent hours tracking down substances in the lab. Keeping track of dozens of compounds is made especially stressful by lab design that doesn’t account for Gabriel’s attention-deficit/hyperactivity disorder (ADHD).

Though their personal experiences in the lab were different, both Sara and Gabriel came to the same conclusion: research labs are seriously lacking when it comes to accessibility. It’s a pervasive problem not limited to one institution or type of disability. Read more to see how they decided to collaborate with LabVoice—a digital lab assistant platform—to develop a tool tailored to their needs.

Camino Financial is using AI technology to loan with empathy

Camino Financial is dedicated to helping small businesses grow to reach their potential. Leading with empathy, this family-founded, financial technology platform startup aims to bring affordable credit to under-banked Latinx micro businesses, and empower underrepresented communities to build generational wealth.

Learn about how Camino Financial is using technology to do good and invest in minority-owned businesses in the following video and its companion blog post.

How Amberdata builds on AWS to unify digital assets for institutions

With 16% of US adults using cryptocurrency—whose market value sat just above $2 trillion earlier this year—and the US government unveiling a plan to regulate digital assets (e.g., cryptocurrency market, blockchain, and decentralized finance data), it is understandable that institutions see digital asset data (e.g., cryptocurrency market, blockchain, and decentralized finance data) as “mission critical” to their success.

Read the blog post to see how web3 startup Amberdata—who recently closed a $30 million series B funding round—is taking the value of informed decision-making and scaling it for digital asset data.

The Amberdata platform unifies digital asset data to support the needs of financial institutions.

How Ergatta partnered with AWS to develop and launch the future of game-based fitness

Sometimes it feels like we’re all searching for the same, elusive thing: a fun workout. Luckily, some people are not just looking for it—they’re creating it. “My other co-founders and I wondered why we had so much trouble making a consistent fitness routine,” says Prasanna Swaminathan, co-founder and CTO of Ergatta. “So we thought back to what made us fit growing up. And that was playing sports—it was having fun. We felt being more like a game removed the need for having more of a structure there.”

Watch the video and read the companion blog post to learn more about the co-founders’ journey and how they’re build the future of game-based fitness.

Cost optimization

Extend your runway by turning off AWS resources when not in use

Think of the last time you forgot to turn off the lights in an empty room; nobody needed the lights on, yet you ended up paying for it. This applies to your cloud resources as well. Fortunately, AWS empowers you to be in control of your spend by giving you not just full visibility on where your spend is going, but also the option to turn off resources when you don’t need them.

AWS offers more than 200 fully featured services on a pay-as-you-go pricing model. This means you pay only for the services you need, and once you stop using them, there are no additional costs or termination fees. This flexibility enables startups to experiment and bring products to market faster than ever before. And this agility does not have to come at the expense of increased costs.

This blog post shows you a few approaches that early stage startups can implement easily in order to get the most out of AWS in a cost-effective manner.

How to prevent unexpected costs for startups while building with AWS

We created this how-to guide to make sure your startup doesn’t end up spending thousands of dollars due to a spike that could have been prevented through monitoring or an alarm. We cover best practices for new and existing AWS accounts when it comes to fundamental security, monitoring, and cost management. We’ll also get into the nitty gritty when it comes to proactively setting up alerts on anomalous usage of services due to over provisioning of services and/or misconfiguration. By following these four recommendations, you can extend your runway and create a long-term strategy for cost management.

Why every startup should set up a budget — and how AWS Budgets makes it easy

As a startup, chances are you’re prioritizing speed to build fast and get your product onto the market as soon as possible. While being laser-focused on your product is essential, it also means it’s easy to overlook your AWS spend, especially if you’re running off credit programs like AWS Activate. You might also have team members wearing many different hats in their roles, making it difficult to spare headcount or attention toward managing costs.

Although monitoring costs on AWS might seem like an arduous task (and make it tempting to ignore the process altogether), it doesn’t have to be. Read on to find out how with AWS Budgets, it takes just a few minutes to set up a budget, which can help you catch surprise bills before they happen. You’ll also be able to monitor costs and usage over time, allowing you to optimize your monthly bill and maximize usage of the perpetual AWS free tier once you’ve transitioned off of credits.

Tech. tips

Should startups use infrastructure as code (IaC)?

When startups opt for “easier,” manual solutions that aren’t reproducible, only the individual who creates the solution has an understanding of its configurations. This leads to configuration drifts — an environment where running your workloads in your infrastructure changes over time due to manual changes or other updates. Drifts can cause your organization time and stress, not to mention monetary loss, even over minor issues.

In this blog post, learn why delaying the implementation of IaC makes it more challenging to deliver new features and fixes to users, and why it takes longer to scale your resources as your user base grows.

Four simple steps to classify your data and secure your startup

A data classification process allows you to distinguish between confidential data and data intended for public consumption, and lets you handle each set accordingly. A startup that has categorized its data can operate more efficiently and more confidently navigate compliance with laws such as the European Union’s General Data Protection Regulation (GDPR) and the California Consumer Protection Act (CCPA).

Understanding your data types and their sensitivity levels ensures that your startup stays ahead of unintended data use or disclosures and satisfies compliance requirements.

To get you started, this post provides four simple steps to simplify and automate the data classification process for your startup.

Business support and advice

Demystifying startup jargon

Entering into the entrepreneur space can be daunting, especially for entrepreneurs who don’t have a grounding in funding, finance, and investment. Read on to get a primer on common terms that will help you run the gauntlet of investment and finance. From non-dilutive capital to uncapped convertibles, shareholder agreements and dry powder, there’s a lot to learn!

The 10 Cs super-successful startup founders have in common

The qualities that make a successful tech startup founder are a complex, even mystical, blend of traits. It’s true there are many other factors which determine whether or not someone will hit the big time, but no matter what the sector, product, or service, we’ve found all super-star founders share key characteristics and attitudes. Here are our top 10 Cs that super-successful tech startup founders all have in common.

Accelerate your startup today

Whether you’re scaling your startup to be the next not-so-hidden gem (unicorn, anyone?) or you’re ready to start building, AWS Startups is here to help you to prove what’s possible. We’re looking forward to seeing what you dream up in 2023!

Ready for the next step? Check out AWS Activate for free tools, credits, and resources that can accelerate your startup at every stage.

Building application-specific blockchains with AWS on Avalanche

Emin Gün Sirer — Wed, 11 Jan 2023 18:00:09 +0000

Imagine a world where transactions—selling a house, trading an asset, paying out a claim—finalize in seconds or less. Contract, payment, and receipts are all tied together in a single action, with the history of the trade permanently logged. The transfer does not require third-party involvement, bypassing centralized middlemen, such as banks, brokers, and agents.

This world is built on blockchain, decentralized and unchangeable digital ledgers that maintain a growing list of transactions stored in “blocks.” Mainstream companies are rapidly embracing blockchains to optimize their costs, while providing faster, safer, more transparent products to their customers.

A recent Alchemy Web3 development report notes that 36% of all existing smart contracts were deployed in 2022. Per Gartner, the business value generated by blockchain is expected to grow from $176 billion in 2025 to $3.1 trillion in 2030.

Realizing the value of academic research and founding Ava Labs

A digital disruptor fueling this mainstream growth of blockchain is Ava Labs, the startup responsible for launching the Avalanche blockchain platform and a primary contributor to innovation happening on the chain. Founded in 2018, Ava Labs is helmed by a team including co-founder and chief executive officer (CEO) Emin Gün Sirer (whose preferred name is Gün) and president John Wu.

“The process that took us to Avalanche started in 2006 and took many years of research in peer-to-peer systems and self-organizing systems that pre-dates Bitcoin and others,” explains Gün, describing his 20 years as a leading professor of distributed systems at Cornell University.

John explains, “Our mission is not just to play in the world of crypto; we want to tokenize all the world’s financial assets to create a better system for everyone.”

The Ava Labs team at Avalanche Summit 2022 in Barcelona, Spain.

Tokenization converts something of value into digital units of asset ownership. Digitized assets can:

Eliminate the inefficiencies of traditional transactions, such as paperwork and human capital
Send easily and quickly
Create greater fractionalization opportunities (division of an asset so more people can participate in ownership) and market liquidity (the ease with which an asset can be bought or sold)

Digitized assets also remove geographical and systemic barriers to access, such as the third-party incumbents who own traditional finance processes.

“It’s time to democratize finance entirely,” says Gün. “You, me, and everybody else can be on an equal footing with the bankers. Scales differ, but the opportunities open to us must be identical.”

Blazing a trail in web3 with novel blockchain infrastructure

Launched in September 2020, Avalanche started a new era for blockchains with near-instant transaction finality. Today, 500+ apps are built atop Avalanche, which is an eco-friendly protocol with minimal energy usage. These apps span the web3 universe, including projects in decentralized finance, blockchain gaming, NFTs, and enterprise use cases. Ava Labs’ work addresses three longstanding challenges in the blockchain space:

Flexibility. The rigidity of monolithic, or single, blockchains can lead to a lack of use-case fit for apps, as well as congestion which results in higher fees for users.
Adaptability. Avalanche allows users to create application-specific blockchains (Figure 1), called “Subnets.” Subnets are customizable and improve the speed and scalability of decentralized applications (“dApps”) hosted on the blockchain.
Scalability. A blockchain platform on which others can build needs to support thousands of transactions per second and have rapid finality.

Launch fast, reliable nodes with AWS to validate Avalanche and its subnets

“Instead of a single chain to rule them all, Avalanche is more like an umbrella where anyone can create their own chain,” says Gün. “These chains isolate activity away from the main network: a load spike on my chain shall not affect the performance of your chain, nor will my fee activity or congestion affect your chain.”

The critical technology behind Subnets is the Avalanche consensus protocol. A consensus protocol ensures the digital ledger is constructed in an orderly fashion and that the nodes sustaining the ledger agree on its state.

There are three principal approaches to consensus protocols:

Proof of work. “This control mechanism was pioneered by Bitcoin, by Satoshi Nakamoto, and is a brilliant idea to use mining to create the chain. The nice thing is that it’s open and anyone can participate,” says Gün. “The issues are scaling and that it consumes a lot of energy.”
Proof of stake (classical or signature accrual). “These are more energy efficient, but they’re not as decentralized,” Gün explains, “Only a few hundred participants can participate in each decision at a time. That’s the downside compared to proof-of-work protocols.”
Avalanche. Per Gün, “Avalanche combines the best of the other two mechanisms: It doesn’t need mining: it’s efficient, it’s fast; it’s decentralized. Avalanche uses proof of stake, but its method of random sub-sampling to achieve consensus is different.”

In addition to the Avalanche consensus protocol, “We are the team that has brought the most recent scientific advances to blockchains. With our platform, with our bridge—which is the biggest bridge in crypto—and with our Core wallet and our new marketplace, Enclave,” explains John. “There is simply no other team that has pioneered as much web3 technology as we have.”

Ava Labs teams with AWS to accelerate web3 adoption

To support its technology, Ava Labs built a serverless architecture and is all-in on AWS. AWS’ cloud solutions, global presence, and proactive partnership are “absolutely essential” in the Ava Labs march to accelerate web3 adoption.

AWS cloud solutions for scalable blockchain infrastructure

Avalanche uses AWS solutions such as Amazon Elastic Compute Cloud (Amazon EC2) to deliver consistent millisecond performance in order to confirm transactions instantly and process thousands of transactions per second.

“It has been a huge, huge boon to all of our developers to be able to spin up nodes on the fly, spin up test networks on the fly, using AWS,” says Gün.

He continues, “Our processes for spinning up Subnets relies on dynamically-provisioned AWS instances so that people can easily, quickly launch their own chain in a couple of clicks.” With AWS global footprint, Avalanche is able to launch customized blockchains, private and public, in seconds to scale with demand.

Ava Labs also participates in AWS Activate, a free program that helps startups to build and scale by offering its members AWS credits, resources, tools, and expert advice. John explains, “AWS Activate is an excellent resource for developers to bring applications to the masses with Avalanche’s trailblazing speed, security, and scalability at their core.”

AWS Marketplace makes it easy for global users to launch a validator node

As a result of AWS’ global infrastructure and commitment to compliance, it is simple for global users to launch an Avalanche Validator Node from the AWS Marketplace. “Number one on my list of how AWS helps us is by allowing customers to launch validator nodes out of whatever legal jurisdiction makes sense for them,” Gün explains.

Ava Labs plans to add subnet deployment as a managed service to the AWS Marketplace in 2023.

John explains, “AWS is so valuable as it streamlines developers being able to go from zero to in-production with dedicated infrastructure. This is especially important to early stage and startup Web3 projects who really benefit from that blockchain-as-a-service model.”

John Wu speaking at Avalanche Summit.

Building the future of web3 with Ava Labs and AWS

In its commitment to providing blockchain execution environments to great developers around the world, Ava Labs is partnering with AWS to host entrepreneur- and developer-focused events.

These events give web3 developers and researchers the opportunity to learn from industry experts and collaborate with their peers to maximize success for their respective startups and products. Educational curriculums include topics such as product market fit, go-to-market strategy, protocol and dApp design best practices, Avalanche-specific technical guidance, and more. Often, these events will feature competitions judged by leading venture capitalists and web3 professionals, ensuring that top-tier individuals and teams find continued support after completion of the program.

“We’re looking for great talent in the developer community who want to make the jump to web3,” John says.

“One of the best things about blockchain is that if you want to develop in this space, it’s a lot easier to start today than when Gün started,” explains John. “Developers can focus on the app they want to build on a Subnet, and they don’t need to do the whole laborious architecture process over again.”

To developers in web3, Gün advises “Part of our success story was to keep the vision incredibly broad, instead of specializing, and to ensure what we are doing ties into the societal problems of the day.”

“Ventures can succeed and fail. At the end of the day, you want to go home and say, ‘I worked on something meaningful today,’” says Gün. “If you tackle something of importance to people and remain true to the things that you know to be good for society at large, you will be fulfilled.”

Shining a spotlight on startup solutions with speedy market entry

Ina Stuve — Thu, 15 Dec 2022 19:30:05 +0000

Gartner defines a go-to-market (GTM) strategy as one that meticulously outlines how a company connects with customers and provides its products and services. This strategy focuses on smart tactics, intelligent buyer journeys, and the effective use of technology. An agile and capable GTM framework enables your startup to build on the foundations of your original idea to create a stable and functional organization that is capable of sustainable growth. In short, a robust GTM strategy can give your startup a serious advantage over competitors.

Here we showcase three winners in the GTM Innovator category of the Amazon Web Services (AWS) Software Startups Awards. These super-charged startups have won customers and secured partnerships through leveraging AWS programs. They’ve also adapted AWS technology and tools to suit their specific needs, and created solutions and platforms to transform how they interact with consumers and enterprises.

Read on to hear more about why we chose Gold medal winner Snyk, Silver medalist Qudini and Bronze medal winner Global Processing Services (GPS) and why each brings something unique to the digital table.

Snyk meets developers where they are

Boston-based cyber security unicorn Snyk is a high-growth startup that has differentiated itself as a GTM innovator by finding new points of entry into target enterprises, using a combination of freemium product and carefully targeted content and SEO. They saw a 2.5x increase in annual recurring revenue (ARR) in 2021 and launched their full developer security platform in the same year.

They used this growth as an opportunity to further their product offering and drive competencies and accreditations, achieving the AWS Security Competency Award, signing a multi-year Amazon Web Services Enterprise Discount Program (EDP), launching two native integrations in the console (AWS CodePipeline and Amazon Inspector), and doubling AWS Marketplace revenue every quarter.

For their GTM strategy, Snyk took an innovative approach to application security. Their developers form a core part of the security solution, so they’re able to develop securely without losing agility or speed. To achieve this, Snyk developed several AWS integrations across the application stack. This way, developers can take advantage of Snyk’s automated security controls from wherever they’re working—from the integrated development environment (IDE) to source control tooling to continuous integration and delivery (CI/CD) pipelines. “We are the only security vendor to have built a first-party integration into CodePipeline so users can scan their open source files for security risks without leaving the AWS console,” says David Lugo, senior manager of partner marketing at Snyk. “This drastically reduces mean-time-to-fix and improves security and agility.”

To further its GTM strategy, Snyk aimed to make it easier for application security and engineering professionals to find and deploy its security solution by offering its tools on AWS Marketplace. Snyk also partnered with Tackle and its cloud marketplace platform, built on AWS Lambda, to provide a zero-engineering approach to listing, integrating, and managing everything the Snyk team needs to sell successfully. This is a key element of Synk’s GTM strategy: recognizing security isn’t just for security teams but also for developers, DevOps teams, and architects within an organization.

“Since we launched in the AWS Marketplace, we closed more than 80 deals in 2021 alone, leveraging both direct and private offers as well as Consulting Partner Private Offers (CPPO). Our revenue has grown dramatically as a result of these two partnerships—AWS and Tackle,” says David. “We also created a Snyk and AWS Developer Security Operations workshop and AWS Quick Start guides to help customers onboard faster with AWS and Snyk, specifically for those using Amazon Elastic Kubernetes Service (Amazon EKS).

How Qudini pivoted and won Silver

Qudini is a smart retail queue management system that allows organizations to drive sales and loyalty across stores and websites. The London-based company had established its position in the global queue management system market by 2020 but when the pandemic hit, it stepped up its game to offer a full “retail choreography” platform, which allows retailers to manage in store queues and offer appointments and event bookings. Its clients include Asda, Primark, and TK Maxx in the UK and Williams Sonoma, East West Bank, and Bass Pro Shops in the US. Qudini has raised £4.5 million (approximately $5.6 million) through seed and Series A funding and in 2021 saw a year-on-year growth of 106%.

“We are now built entirely on AWS, and this has allowed us to expand internationally. We can rapidly roll out services in different countries while still meeting local data requirements,” says Fraser Hardy, CTO and co-founder of Qudini. “We have a proven process to spin up new installations in any AWS Region required by any customer, with our fastest roll-out expanding a customer from zero stores to 500 in less than two months. This much-needed functionality has transformed our business. When we started out, it was a challenge to deliver our service in a country where we didn’t already have a presence.”

The system can now roll out into new Regions seamlessly, spinning up new installations on demand. Qudini has several customers with global branches that can use the solution across multiple Regions. Currently, Qudini has services in five Regions with a AWS usage of $45,000 per month and is working with the AWS partnership team to become an ISV Accelerate partner—Qudini has already shared more than 10 go-live projects and has more in the pipeline.

“We have a fully verified solution that has passed a foundational technical review in record time due to our previous work, and we are in the process of providing our service on AWS Marketplace,” says Fraser. “We hope to complete this process early in 2022 alongside our Retail Competency certification.”

“Use the tools provided to avoid reinventing the wheel and focus on delivering the value of your product,” says Fraser when asked what startups can do to build robust GTM strategies.

Bronze is awarded to Global Processing Services

GPS is a payment processing solution designed for the payment and Fintech sectors. The company has been in operation since 2007 and is committed to accelerating the delivery of better financial experiences globally.

“After a proven track record using AWS for development teams and building out our continuous integration and deployment solution, our IT team has further expanded our usage of the AWS product suite to meet the strategic global ambitions of our company,” says Hannah Taylor, head of marketing at GPS. “Our investment in automation and DevOps has accelerated our developer advocacy program, allowing our current and future customers to test and ‘try before you buy’ by building a comprehensive on-demand program.”

GPS worked closely with AWS to re-architect its platform to create a solution that’s agile and capable of meeting its strategic global ambitions and those of its customers. The company realized that in order to scale intelligently and achieve a competitive advantage, including its GTM strategy, it needed to establish a robust partnership with AWS built on mutual trust. The startup has now built a new GPS environment aligned with AWS best practices using AWS tooling. The first European cloud offering passed the stringent PCI-DSS AoC level 1 audit to operate in September 2021.

“Our fast growth global expansion wouldn’t be possible without AWS,” says Hannah. “We are continuing to expand our horizons using native AWS technologies such as Amazon API Gateway. We also use the AWS Marketplace to quickly implement solutions and maintain security confidence using trusted products such as Palo Alto Networks.”

GPS will continue to focus on further development with AWS using Amazon EKS and Amazon Relational Database Service (Amazon RDS) for microservice roll-outs across the globe. In addition, the company plans to continue working closely with an AWS Lighthouse project to fulfill the needs for fully compliant PCI DSS Hardware Security Modules.

“The full AWS ecosystem allows us to construct what we need, when we need it,” concludes Taylor. “Without Amazon Elastic Compute Cloud (Amazon EC2), we couldn’t operate. AWS Control Tower allows us to keep secure, and Amazon RDS allows us to quickly implement a new postgres with ease. Without things like AWS Transit Gateway, we wouldn’t be able to join secure components globally.”

Conclusion

The AWS Software Startup Awards recognizes innovative startups and entrepreneurs across several key categories, including Sustainability, Rocketship, Founder of the Year and Rising Star.

SaaS Founder Series: Epistemix on how software simulations can create more empathetic decision-making

Jessica Koh — Mon, 12 Dec 2022 21:42:32 +0000

Sometimes companies that offer software-as-a-service (SaaS) need help creating scalable infrastructure so they can deliver their product to a larger audience. AWS SaaS Factory invited Epistemix’s chief executive officer (CEO) and co-founder John Cordier to share how the company is using Amazon Web Services (AWS) to scale, and how their software is helping data scientists make more informed, empathetic decisions based on real and synthetic data through machine learning models.

Humble beginnings

In 2000, world-leading epidemiologist Donald S. Burke was heading infectious disease research for the US military when he realized that decision makers didn’t have the tools they needed to understand what the future of an epidemic might look like.

To address this gap in the market, Burke, Cordier, and John Grefenstette founded Epistemix, an agent-based modeling platform paired with a synthetic, statistically accurate population of the United States.

With Epistemix’s software, data scientists can gain insights by running experiments—controlling variables like climate or economic changes—in a statistically accurate virtual world, exploring how large human populations might behave given different circumstances, as shown in Figure 1.

Figure 1. Users can immediately start running models on the platform’s synthetic population of the entire US

Epistemix started small, incubating the software within the Graduate School of Public Health at the University of Pittsburgh. They also received funding from the National Institutes of Health and the Bill & Melinda Gates Foundation.

But when the three co-founders realized that their platform could help solve social problems beyond the realm of infectious diseases, they spun it out into a company—and Epistemix, in its current form, was born.

Expanding beyond infectious disease response

“When we initially founded the company, our goal was to change the way that public health would be practiced everywhere in the world,” says Cordier. “And public health in the United States is not prioritized. Every single person became a witness to that over the last two and a half years.”

During the COVID-19 pandemic, the Epistemix team used their software to help clients in the global events industry—including CES, IMTS, the Javits Center, and other organizations—make better decisions about how and when to bring people back together.

While they were proud of this work, the Epistemix team soon realized that it was time to scale. The next step would make their platform accessible to the public, so that external data scientists could use it, vastly expanding the tool’s social impact on the world.

“We’ve been able to simulate different product adoption trajectories for different products depending on what marketing strategy they’re going with,” says Cordier. “We’ve been able to simulate segregation in the housing market (Figure 2) in the US. So, what’s next for us is really getting our software in the hands of data scientists in other areas.”

Figure 2. Simulating housing equity with a classic Schelling model

SaaS journey with the help of AWS and SaaS Factory

Epistemix’s first challenge in sharing their technology was creating a scalable infrastructure that external data scientists and other organizations could use. While their platform itself was already quite robust, they struggled to make the software more user-friendly.

After looking at other cloud providers, they partnered with AWS based on its broad range of services. And when they found themselves ready to launch with external users but were still running into technical challenges, they engaged the AWS SaaS Factory for guidance.

“Relying on the AWS SaaS Factory’s expertise, we’ve been able to create a better user onboarding experience. And through that better onboarding experience, we hope to grow the impact that our company can have,” says Cordier.

“AWS SaaS Factory helped us deliver the tool that we’ve built—with the sole intention of improving population health, equity, and policy decision-making that’s going to impact entire groups of people—to data scientists and other organizations.”

Making large-scale, informed decisions with empathy

“Every single person [in the US] is represented,” says Cordier of the Epistemix platform. “And so, I view what we are doing as giving every single person in the country a voice through data.”

Cordier believes that by sharing their platform, which one can use to see how the effects of many different variables play out, scientists will begin to understand how interconnected humans really are.

“When somebody is able to think differently about solving a problem, you can get to some really compelling solutions and a deeper level of understanding rather than what’s coming out of a very linear Excel spreadsheet,” says Cordier.

With this in mind, he believes that Epistemix has the power to change the way people think about solving a problem.

“And if we can have perspectives that are rooted around entire populations, I believe that creates more empathetic leadership, because you’re thinking about the population that your decision is going to impact.”

About AWS SaaS Factory

AWS SaaS Factory can help startups at any stage of their SaaS journey. Whether you’re looking to build new products, migrate existing applications, or optimize SaaS solutions on AWS, we can help. Visit the AWS SaaS Factory Insights Hub to discover more technical and business content and best practices.

Reach out to your account representative to inquire about engagement models and to work with the AWS SaaS Factory team.

Alloy’s global identity decisioning platform, built on AWS

Charles Hearn — Fri, 09 Dec 2022 18:52:44 +0000

Entrepreneurs know all about risk versus reward: short-term trade-offs, long-term gains, and some decisions so complex that solving one conflict raises another. This is particularly true in the financial industry, where success is built on choosing the right clients.

Alloy co-founders Laura Spiekerman (president), Tommy Nicolas (chief executive officer CEO), and Charles Hearn (chief technology officer CTO), saw these challenges as a business opportunity.

In 2015 they founded Alloy, a unicorn fintech startup whose global identity decisioning platform helps banks and fintech companies automate their decisions for onboarding, transaction monitoring, and credit underwriting.

Alloy combines the use of traditional data sources (such as credit scores) with newer, alternative data sources, such as cash flow data, to provide a complete and more accurate picture of each customer.

“We tell you which aspects of a client’s identity match, and which don’t match, or match in one record source but not another, and here’s an interesting flag that we should show you,” explains Charles.

Effective identity decisioning helps companies to:

Minimize financial risk
Maximize information security to avoid fraud
Pave a path to smooth and seamless client onboarding

Working backwards from the problem

With experience in startups and payments processing, the trio of co-founders is well-versed in the identity verification problem. Alloy is built from their collective experience, determination, and—as the founders readily admit—a fair bit of luck.

At the payment processing startup where Laura, Charles, and Tommy met, “We saw firsthand that the answer [to identity decisioning] was for each company to make relationships with data vendors and then have engineers integrate them. It wasn’t core to the business, but it was the thing you had to build,” says Charles. “Businesses were losing time and resources by building the identity-decisioning infrastructure instead of incorporating a ready-made solution.”

Laura, Charles, and Tommy saw the opportunity to create a better solution for companies whose success was dependent on choosing the right clients. They left the payments processing company and founded Alloy.

From left to right: Charles Hearn, Tommy Nicolas, and Laura Spiekerman of Alloy

In the beginning, explains Laura, “Charles was building the product with a couple of engineers, I was having conversations with prospects and customers, and Tommy was implementing the solution and getting product feedback.”

Today, their API-based platform services over 300 banking and fintech clients.

Providing a singular solution

What sets Alloy apart? Quite a lot, actually.

For starters, Alloy clients have the power to make informed decisions.

Transparency into the data within the Alloy platform allows clients to tweak their risk tolerances to determine who to approve or deny. Clients can set mutually dependent parameters of who to flag, all without ever touching a line of code.

The specificity of what you can flag and let through helps Alloy serve a broad range of clients. For instance, a credit card company has a different risk tolerance than a bank or a bitcoin company.

The Alloy platform provides a detailed picture of each identity-decisioning case

Clients can use these features to instantaneously onboard customers who may have otherwise been sent to manual review or declined due to a lack of data or a mismatch between the data and the risks.

“Our product lets you see the outcomes,” explains Laura, “so you can easily see how you’re optimizing for fraud mitigation or for conversion. We give you the results and you can spot where your performance is the best.”

Creating broader access to financial services, while making it safe and easy for the companies involved, is a core component of Alloy’s mission. “We have always taken our position in approving and denying people for essential financial services really seriously,” explains Charles. “The core motivation behind our product is to make those experiences safe and easy to do.”

Building on AWS

One way Alloy is giving their customers an easy-to-use experience is by building a cloud-native product. Charles laughs as he recalls how, in 2015, “I got a lot of questions from our initial clients about, ‘Can you build financial infrastructure on the cloud?’ Today it’s obvious that you can, but at the time only a few companies were doing it.”

For their cloud technology, Alloy is all-in on AWS. Whether that be their core PostgreSQL database that runs on Amazon Aurora; their compute, which runs their containerized applications on Amazon Elastic Kubernetes Service (EKS); or Amazon OpenSearch Service, which Alloy relies on for log analytics and search. Beyond these fundamental technological components, the Alloy teams also use Amazon ElastiCache for their caching needs and Amazon Managed Workflows for Apache Airflow (Amazon MWAA) to orchestrate their data pipelines. They also use AWS machine learning and analytics services to enhance their decisioning and monitoring services.

“We joined very early with AWS. A lot of the products AWS offered made us stand out and made it easier for us to get off the ground, especially for providing security,” says Charles.

Alloy joined the AWS Activate startup program through Techstars which, “was essential to our growth in our first two years,” said Charles. “Through the program, we didn’t have to pay for server costs for a long time and AWS provided an experienced solutions architect to streamline our infrastructure in those earliest days.”

“Using AWS services makes it easier to optimize building your product instead of building all of the things that you think you have to build that are peripheral to it,” says Charles. “That’s the same thing we try to do for our clients—take things off their plate, and handle the things that are not core to their business, so they can focus on their core business objectives.”

Strong collaboration, proactiveness, and openness between Alloy and their AWS account team has led to more effective solution implementation, accelerated progress on key initiatives, and support for launching new products. “AWS is always helping us assess how to optimize our tech stack, especially as we’ve launched new products like transaction monitoring, which required different scalability requirements than our original onboarding product,” explains Charles.

Charles Hearn speaks at the AWS Summit

With opportunities for on-site solution immersion days, product team engagement, solution brainstorming sessions, and strategic sessions to work backwards from a customer goal, it is easy to see how leaning into a relationship with AWS can accelerate startup success.

Going broader and going global

What’s next for this fintech unicorn?

To make financial experiences even more safe and seamless, Alloy expanded its solutions to include transaction monitoring and credit underwriting in 2021.

“We’re creating a unified view of the customer lifecycle: we plug our onboarding system into our transaction monitoring system and our credit underwriting system. This will allow our clients to make more intelligent identity and risk decisions about the customer,” explains Charles. Creating a system of products that talk to each other and feed off of each other’s information will be much more powerful than analyzing smaller pieces of customer information in a vacuum.

Along with broadening their product range, in August of 2022 Alloy expanded internationally. Per Charles, “In part, we’re able to do this because AWS allows us to easily create a data center in Ireland. We’re also creating one in Australia, to more easily serve clients around the world.” To date, clients in 40 countries can access Alloy’s services and they expect the number to grow.

“We’ve seen through the first two or three innings,” says Laura. “We still have a lot more to go.”

Discover four Pakistani startups at the forefront of AI/ML

Eunice Cheng — Tue, 06 Dec 2022 18:27:47 +0000

Amazon Web Services (AWS) and Epiphany joined forces in 2021 to co-curate an artificial intelligence (AI)/machine learning (ML) bootcamp called AI/ML Reactor. AI/ML Reactor is a rigorous 5-week virtual program aimed at driving AI/ML awareness and empowering startups in Pakistan.

We received an overwhelming response for this program. Twenty-five startups were chosen out of 250 startups that applied from all provinces in Pakistan. Participants had access to exclusive master classes, a group tech mentoring session, and one-on-one mentoring sessions with AWS specialists and thought leaders. At the end of the program, they presented their AI/ML solutions to a panel of judges (see Demo Day).

Meet our winners from the 2021 Reactor!

Salesflo – 1st prize

Salesflo is one of Pakistan’s fastest growing software as a service (SaaS) platforms. They build tools to improve in-field sales efficiency for consumer goods.

The Salesflo team embarked on AI/ML Reactor to build Salesflo Airstrike – an image recognition solution that automates the retail merchandizing process. It uses Amazon Rekognition, Amazon SageMaker, Amazon Elastic Compute Cloud (Amazon EC2), and AWS Lambda.

Currently, only 2% of Pakistan’s 700,000 retail outlets are merchandized. This provides growth potential for fast-moving consumer goods (FMCG) companies to make their products available, visible, and engaging to consumers. Salesflo’s winning solution enables a shop to be merchandized, providing greater store coverage, lower costs, and faster availability.

Since winning AI/ML Reactor in 2021, Salesflo has completed successful pilot tests with clients such as Unilever and Friesland Campina. In addition, they have been working on embedding this module within their product Engage to enhance merchandizing and optimize workflows. This past year, Salesflo has been working on optimizing the models that they are using and determining product divisions and pricing categories.

“The program is a step toward fulfilling the potential of the retail industry and bringing innovation to the up-and-coming tech sphere of Pakistan” – Hamza Khan, Head of Data & Engineering, Salesflo

Ozoned Digital Ltd (“Ozoned”) – 2^nd prize

Ozoned is an insurtech startup that aims to digitally transform the insurance value chain. It services multiple stakeholders (insurers, insurance brokers, insurance agents, customers, and others) in the insurance ecosystem.

Ozoned joined the AI/ML Reactor program to develop solutions to overcome digital data entry errors and minimize high survey costs in the automotive insurance space. With Amazon Textract, Ozoned was able to extract data automatically from a customer’s government-issued identification card to activate motor insurance coverage.

To help minimize insurance claims surveyor interaction during the initial stages of a claim, Ozoned used Amazon Rekognition with custom labels. This enabled identification of a vehicle and its damaged parts, reduced costs, and created efficiencies. The model was further trained to assess whether the vehicle damage was a partial or total loss. This assessment helped insurers generate an initial claim estimate for the damages, while reducing time and cost to the insurer.

Most recently, Ozoned has signed up one of the largest insurers in Pakistan – Adamjee General Insurance. Adamjee is deploying Ozoned across their motor insurance operations.

“The AI/ML Reactor program allowed Ozoned Digital to impart AI/ML training to its team and come up with a world class AI/ML technology solution in the insurtech space.” – Nomaan Bashir, co-founder and CEO, Ozoned Digital Ltd

XpertFlow – joint 3^rd prize

XpertFlow is an AI-powered preventative healthcare company founded in 2019. Its mission is to reduce mortality from hospital acquired infections (HAIs) that eventually lead to sepsis.

In Pakistan, approximately 350,000 people are affected by sepsis each year, and more than 275,000 lose their lives. It costs public and private hospitals more than 10 billion PKR ($66 million) per year to treat sepsis. Without timely treatment, sepsis can rapidly lead to tissue damage, organ failure, and death.

During the AWS AI/ML Reactor program, the XpertFlow team used AWS services to forecast sepsis 6 hours before its onset:

Amazon Simple Storage Service (Amazon S3) stores raw time series data and the model artifacts.
Amazon SageMaker’s Data Wrangler and notebook instance handles the entire pre-processing pipeline.
Amazon SageMaker Autopilot trains over 200 different iterations of the model, with different algorithms available on SageMaker. With Amazon SageMaker Autopilot, the team was able to select a model from several generated based on the area under receiver operator curve (AUROC).
An endpoint was generated from that model and connected to Amazon API Gateway using Lambda.
The API was then connected to the demo interface on the team’s deepnostiX

The winning model delivered 97% accuracy in predicting sepsis 6 hours ahead of time. It acts as an early warning system that performs nearly continuous monitoring of patients in ICUs or wards, assigning a risk score based on a patient’s vital signs.

After the AI/ML Reactor program, the sepsis AWS AI model is now fully tested and is being piloted at a few hospitals in Pakistan. XpertFlow has started using SageMaker to estimate blood pressure noninvasively, continuously, and without using arm cuffs. They are currently running trials for this new approach towards calculating blood pressure in a hospital in Islamabad, Pakistan.

“Being a CTO of a young startup, I am always looking out for tech-focused programs. After attending many programs and accelerators during the life of XpertFlow, this has been by far one of the best. The learning the team and I got from this, during a span of 5 weeks, was just incredible. We learned from the best on how to deploy our existing AI/ML solutions on the cloud and supercharge them using AWS AI/ML services. Huge shout to AWS and Epiphany for coming up with one-of-a-kind program, which was missing here in Pakistan.” – Shan Ul Haq, CTO, XpertFlow

Trukkr – joint 3^rd prize

Trukkr provides financial services and technology for logistics in Pakistan. It gives both large and small businesses a comprehensive technology platform to manage and provide all their logistical needs. Trusted by some of the biggest companies in the country, Trukkr saves organizations time and money, while providing them with deep data and powerful insights.

In an effort to improve dropoff rates during sign up (see Figure 1), Trukkr investigated their onboarding process. During the AI/ML Reactor program, Trukkr delivered initial solutions to optimize their onboarding funnel. They extracted data from uploaded documents using Amazon Textract, and used that information to prefill a signup form.

Figure 1. Trukkr’s dropoff rates during sign up

The team was able to make an impact on improving the onboarding funnel, increasing their signup completion rates from 38% to 52%.

“This program helped us in further leveraging AI/ML to provide our customers with intelligent features that help them better manage and optimize their logistics. During the program, regular feedback from AWS expert mentors and the step-by-step architecture review sessions helped speed up the learning process, and deliver meaningful features to our customers.” – Kasra Zunnaiyyer, co-founder and CTO, Trukkr

Ten must-attend re:Invent sessions for Fintech startups

Phoebe Chen — Wed, 23 Nov 2022 19:13:57 +0000

There is a ton to be excited for at this year’s re:Invent! If you haven’t already, be sure to register here. To help you plan, I’ve put together my top 10 sessions for Fintech that you don’t want to miss.

1. Making the right hiring decisions to build impactful products

In this breakout session, you’ll learn from Ali Heron, CTO of Petal, a financial inclusion Fintech startup, about how she doubled the size of Petal’s product, engineering, and analytics teams in just under a year with scalable hiring practices for quick decision-making and driving impact for customers.

2. Combating fraud with identity verification AI workflows

Tune in with our Identity Verification and Fraud Prevention team during this Chalk Talk to learn how AWS can help you automate and streamline identity verification and fraud detection for risk management and customer onboarding, including know your customer (KYC)/know your business (KYB) and anti-money laundering (AML) with artificial intelligence (AI) and machine learning (ML).

3. How Coinbase rapidly builds and scales their NFT marketplace with AWS

In this Chalk Talk, you’ll learn how Coinbase uses the latest innovations on AWS to scale and maximize compute power in the dynamic world of Web3 for their new non-fungible token (NFT) marketplace business.

4. Serverless order fulfillment with Stripe and AWS (sponsored by Stripe)

Hear from Stripe in this Lightning Talk about how you can use their webhooks with Serverless on AWS services such as AWS Lambda to handle key events in your payment workflow and scale your solution.

5. Build human-like customer experiences with conversational AI

Start building fully automated human-like chat experiences for your customers in this workshop using conversational AI with Amazon Lex, Amazon Kendra, and the Amazon Chime SDK.

6. Deploying egress traffic controls in production environments

In this breakout session, you’ll join Robinhood’s engineering team to learn how they implemented AWS Network Firewall to manage network traffic, block threats, and detect anomalous activity on workloads that process sensitive financial data.

7. Identify improper payments with analytics and ML

Learn best practices from this builders’ session on how to build a strong solution to identify improper payments with advanced analytics, ML, and deep learning techniques.

8. Patterns for obtaining insights from sustainability data

In this Chalk Talk, discover how AWS Data Exchange makes it easier for you to find, subscribe to, and use the data needed to support your environmental, social, and governance (ESG) strategies and how to successfully integrate them with AWS architectures.

9. Analyzing use cases and running applications on quantum computers

Tune in to this Chalk Talk to see how the Amazon Quantum Solutions Lab and Goldman Sachs collaborated to deliver a successful quantum computing project by estimating resource use, optimizing computational strategy, and analyzing results.

10. Fintech Networking Reception

As you unwind from a day of re:Invent sessions, join leaders from fellow Fintech startups, financial services enterprises, and the AWS Fintech sales team for a drink (or two!) to network and discuss partnership opportunities.

A few other sessions worth checking out

1. Habits to reduce your cloud costs

Managing costs is a critical topic for startups. Check out this breakout session to learn how you can better optimize your cloud spend and get the most out of AWS.

2. AWS Impact Accelerator: Being the megaphone

Earlier this year, we launched the AWS Impact Accelerator to support underrepresented founders who are building the future. Join this breakout session to hear from program participants and VP of AWS for Startups, Howard Wright, to learn how to lift up and shine a light on the next generation of entrepreneurs.

3. Cloud strategies for equity valuation and investor readiness

In this breakout session, learn how cloud strategy plays into equity valuations and how AWS can help you stay market and investor ready.

See you there!

Hope you’re as excited for this lineup as much as I am! re:Invent is a great opportunity to connect with industry leaders and catch up on all of the topics you’ve wanted to dive deeper into. Be sure to stop by the AWS Startup Loft booth and meet our startup specialists.

Want to know more?

You can find more details on our re:Invent startups page.

How Amberdata builds on AWS to unify digital assets for institutions

Tongtong Gong — Tue, 22 Nov 2022 20:03:32 +0000

Informed decision-making is a valuable process to which many startup founders relate. Whether during product development, assessing market fit, or meeting customer needs, timely and accurate information can be the difference between success or failure.

Web3 startup Amberdata—who recently closed a $30 million series B funding round—is taking the value of informed decision-making and scaling it across an emerging sector: digital asset data.

Digital asset data includes:

Cryptocurrency (crypto) market data, such as live and full historical prices for specific crypto assets.
Blockchain data, which includes records of network data, contracts, events, and logs for every digital asset.
Decentralized finance (DeFi) data, which provides insights ranging from digital wallet earnings and losses, transactions, and balances to DeFi asset level data including trading pairs by volume, price, and time.

With 16% of US adults using cryptocurrency—whose market value sat just above $2 trillion earlier this year—and the US government unveiling a plan to regulate digital assets, it is understandable that institutions see digital asset data as “mission critical” to their success.

Amberdata provides institutions with a unified data integration point from which to glean up-to-date insights about crypto markets, blockchain networks, and decentralized finance.

Becoming the leading provider of digital asset data to financial institutions

The inspiration to found Amberdata in 2017 was the initial coin offering (ICO) hype, when co-founders chief executive officer (CEO) Shawn Douglass, chief operations officer (COO) Tongtong Gong, and chief technology officer (CTO) Joanes Espanol—who are all former colleagues—became intrigued by the crypto world.

Tongtong Gong, COO and co-founder of Amberdata

COO Tongtong Gong recalls, “We started reading a lot about Ethereum, smart contracts, and tokenization,” which led to, “going down the rabbit hole of wanting to learn more about the crypto industry.”

She and her co-founders, “realized there was not enough data and tooling to enable access to clean, normalized, easy-to-ingest and understand data to help people in the digital asset space.”

They began building a platform to provide institutions with the tools and data necessary to adopt crypto offerings and meet clients’ demands for an alternative asset class.

“When you start a company, all you have are your co-founders, six slides, and a dream, right?” says Tongtong with a laugh.

Fast-forward five years and Amberdata is now the leading provider of digital asset data to financial institutions, who use the platform to make informed decisions while accelerating their time to market.

The Amberdata platform (Figure 1) allows clients to leverage a broad, granular dataset to offer digital asset product and services to their customers, while sidestepping the significant hurdle of creating and managing an in-house data pipeline:

Figure 1. The Amberdata platform unifies digital asset data to support the needs of financial institutions.

Building their success with AWS

Choosing Amazon Web Services (AWS) to host Amberdata’s cloud-native platform, “was the easiest decision for us to make when starting the company,” says Tongtong. Along with the three co-founders’ prior experience building on the AWS tech stack, they also chose AWS because, “It’s easy to hire developers when AWS is your platform because so many of them already have the knowledge to use it.”

Amberdata joined AWS Activate, a free program designed for startups and early stage entrepreneurs that offers the resources needed to get started on AWS.

“AWS gave us generous credits to kickstart our proof of concept. As a bootstrapped startup, the credit really helped us to reach our first MVP.”

With clients in the highly-regulated financial industry, Amberdata sought a cloud solution that was secure and reliable, while also providing efficiency and optimizing their costs. As part of the Activate program, “We had help from an AWS solutions architect and went through an AWS Well-Architected review to make sure what we built was secure, reliable, and performance- and cost-efficient.”

Today Amberdata is, “completely built on AWS. From ingesting the data, to processing it, extract-transfer-load (ETL), and to serving the data—whether that be with Amazon Elastic Compute Cloud (Amazon EC2), Amazon Simple Storage Service (Amazon S3), or different Amazon Elastic Container Service (Amazon ECS) containers. I think the commitment is still true: The AWS team really helps us to use every single lever to make sure we are building an efficient, secure, and scalable product.”

Advice to startup founders

To aspiring founders, Tongtong shares that, “Starting a new company as an entrepreneur, I feel like I get to define my own role. I’m interested in so many things, and felt like I had hit the ceiling after becoming a vice president of engineering. Being a founder allows me to get my hands dirty, to be a free agent, and choose my own path.”

“Being a founder is like marrying the best of two worlds. Utilizing our past experience—15 years of enterprise software experience and data experience—in this new crypto space,” says Tongtong.

The number one lesson she’s learned as a founder is that, “You’ve got to have a customer-centric view. Whether it’s how you operate, how you treat your customers, or how you solve problems—focus on the customer’s need. If you solve that problem, the revenue will come. Everything else will follow because you’ve provided value.”

She also advises, “Never underestimate yourself. You are capable of doing so much more than you think you can do. Whatever it is, just go for it. Ask for help—more often then not, people are willing to help you if you just ask the question.”

Looking to the future for Web3 and Amberdata

Amberdata’s next steps include building more analytics insights into digital asset data.

Additionally, they are expanding their blockchain offerings with a goal to become, “the de facto best—the most comprehensive, deepest, broadest data analytics company that gives everyone access to the data.”

The Amberdata team

Providing institutions with a path to modern asset offerings is part of Amberdata’s bigger picture that includes, “democratizing financial access to historically unbanked populations,” and, “pushing web3 adoption forward.”

“I’m passionate about crypto because you don’t often see a new asset class that represents so much opportunity,” says Tongtong.

Attending re:Invent 2022? Be sure to check out Tongtong Gong at the Women’s Panel session STP217.

Fast, reliable, and cost-efficient builds/tests at scale with EngFlow Remote Execution on AWS

Christian Mueller — Mon, 21 Nov 2022 17:37:36 +0000

Even as modern engineering organizations adopt microservice architectures and decompose their monolithic applications, large and complex code bases are common. More code leads to longer build/test cycles, which degrades developer productivity and increases cost. In addition, finite continuous integration (CI) compute budgets cause execution queuing, which further saps engineering productivity.

EngFlow helps modern organizations improve their build and test cycles to maximize productivity of software development teams through Remote Execution and Caching.

To provide performant, reliable, and cost-effective solution to its customers, EngFlow followed the best practices from the AWS Well-Architected Framework. In this post, we’ll focus on the practices that helped improve the price-performance ratio and improve availability.

EngFlow Remote Execution service’s architecture

EngFlow customers interact with the Remote Execution service over a secured and private channel, protected on several layers using different network technologies like Network Load Balancer and virtual private cloud (VPC) endpoints, subnets, and security groups (see Figure 1).

Schedulers, which run on Amazon Elastic Compute Cloud (Amazon EC2) instances with Amazon Elastic Block Store (Amazon EBS) volumes, divide each build/test request into independent parts and place the individual build/test jobs on existing worker instances, which fulfill the build/test job compute/memory requirements. AWS Auto Scaling is used to provide self-healing capabilities by maintaining a fixed number of running scheduler instances. Failed Amazon EC2 scheduler instances are replaced automatically without the need for human intervention.

EngFlow’s Remote Execution software distributes build/test actions across hundreds or even thousands of Worker instances, which all run on EC2 instances with mounted EBS volumes. Here, EngFlow uses AWS Auto Scaling to scale the required compute capacity in and out based on demand, minimizing waste and maximizing utilization. To support different customer needs and offer a cost-efficient solution, Worker instances can run on EC2 On-Demand or Spot Instances. EngFlow provides the ability to fall back to On-Demand instances for greater reliability.

Remote Execution pairs with EngFlow’s Remote Caching solution to prevent work duplication. By caching all build/test artifacts on local storage performed by an organization, EngFlow allows teams to download previous build/test artifacts rather than to re-execute the build/test themselves. This shortens the build/test time and decreases the cost per build/test. As instances in the cloud come and go, the build/test artifacts are durably and cost-efficiently synced to Amazon Simple Storage Service (Amazon S3) for persistent access.

To support operations when deployed into a customer account, EngFlow’s Remote Execution solution uses AWS standard services like Amazon CloudWatch Logs, Metrics, and Alarms.

Figure 1. EngFlow Remote Execution solution architecture on AWS

Let’s take a deeper look at how EngFlow enables fast, reliable, cost-efficient builds on AWS and how they identified areas to improve their Remote Execution solution when working through their Well-Architected review.

Performance

Many EngFlow customers consider build/test speed as one of EngFlow’s primary value propositions. It translates to increased developer productivity, faster time-to-market and the ability to run more experiments at the same time, just to name a few. Running a CI build/test in the cloud and using distributed compute nodes from the latest and most powerful CPU generation instead of a local workstation helped customers like Blue River Technology achieve a 9 times performance gain for their CI build/test.

EngFlow engineers help their customers select the most performant and cost-efficient EC2 instance that best fit their unique requirements, like the latest generation AWS Graviton3 processors. They use the latest EBS gp3 volumes, which provide the latest general purpose SSD volumes generation to customers to ensure builds/tests are as fast as possible. This was possible by analyzing the unique customer workload, leveraging standard and custom CloudWatch metrics.

In addition, the ability to fine-tune the AWS Auto Scaling configuration for Worker instances helps EngFlow find the optimal balance between capacity and cost so their customers always having enough compute and storage capacity to start a new scheduled build/test immediately, while not wasting money on idle resources.

Reliability

EngFlow has been following AWS best practices for providing a reliable service from the start to run time-critical build/tests for thousands of developers.

Their AWS Auto Scaling configuration spans across three AWS Availability Zones in the selected AWS Region to scale based on a customer’s demand, withstand a local service disruption, and provide self-healing capabilities by replacing failed Amazon EC2 instances automatically. Amazon S3 complements the solution as a durable and highly available storage service for build/test artifacts.

During the Well-Architected review, EngFlow discovered an area to improve the reliability of their service. Amazon EC2 Auto Scaling uses termination policies to determine which instances it terminates first during scale-in events. Termination policies define the termination criteria that are used by Auto Scaling when choosing which instances to terminate.

By default, the termination policy selects the Availability Zone with the most instances. It terminates the instance that was launched from the oldest launch template or launch configuration. If the instances were launched from the same launch template or launch configuration, Amazon EC2 Auto Scaling selects the instance that is closest to the next billing hour and terminates it.

EngFlow observed that this default termination policy sometimes terminated EC2 instances that were in the middle of executing valuable work for our customers. While EngFlow Remote Execution automatically retried the job, it led to longer build/test executions, higher cost, and customer inquiries. By creating their own custom termination policy in combination with using instance scale-in protection, EngFlow was able to increase their reliability and improve efficiency.

Cost efficiency

EngFlow customers run CI builds/tests at large scale, using hundreds or even thousands of EC2 instances concurrently at peak time. This workload makes a sweet spot to look for cost optimization opportunities on behalf of our customers.

To save up to 90% on compute costs compared to on-demand, in our Well-Architected review we discussed with EngFlow the ability to enable Spot Instances for their Remote Execution, based on Amazon EC2 Spot Fleet.

“This feature is a huge opportunity for our customers and us to become more cost-efficient when running in AWS. We loved implementing it. By design, our Worker instances are stateless and handle spontaneous termination with a robust retry mechanism. AWS Spot Instances were a perfect fit, saving EngFlow and our customers 70% on compute costs on average.” – Yannic Bonenberger, Engineer at EngFlow explained.

While working on the Spot Instance integration, EngFlow also improved the overall EC2 instance resource utilization by improving our Worker instance scheduling algorithm, resulting in additional cost savings for our customers.

After EngFlow integrated Spot Instances, they listened to the Spot Instance interruption notices, which are issued two minutes before Amazon EC2 stops or terminates your Spot Instance. With this, they could avoid scheduling new build/test jobs on instances that will get reclaimed soon. Similar to this, the AWS Solution Architect also recommended listening to the EC2 instance rebalance recommendations to avoid scheduling new build/test jobs on instances with a high likelihood of getting reclaimed soon as well. After implementing these recommendations, EngFlow customers observed fewer build/test job retries because of premature EC2 instance terminations.

A promising approach to modern builds/tests

Large, nimble technology organizations such as social media platforms, short-term travel marketplaces, and auto manufacturers rely on EngFlow’s platform to keep engineers in flow and maintain the necessary agility for modern software development. AWS is at the core of EngFlow’s success, giving them flexible architecture and cost efficiency, which directly translates into competitive advantage for end customers.

“It’s incredible how much power you get at your fingertips. This is the first time in history that you can get 1,000 machines as a single developer and try something out at scale.” – EngFlow’s CTO Ulf Adams states.

Varjo is pushing the boundaries of mixed reality with the help of AWS

Jussi Mäkinen — Fri, 18 Nov 2022 13:29:08 +0000

Most of us are familiar with how virtual reality (VR) can transport us to a make-believe realm. But how can it help us tangibly improve our physical world? For the past six years, the Helsinki-based VR/extended reality (XR) startup Varjo has been creating professional hardware, software, and services to help product designers develop consumer electronics that don’t yet exist.

Varjo’s many customers include Boeing, whom they partnered with to create astronaut training simulations, and Volvo, whom they worked with to make the first-ever mixed reality driving experience. Volvo car designers now use this technology to virtually evaluate prototypes, safety features, and designs—saving both time and money.

A person using the Varjo virtual reality and extended reality hardware.

Boeing is one of the many companies using this technology to break new ground. While astronauts have been relying on immersive technology to train for space missions for years, the quality of the resolution was never high enough to allow them to prepare for all safety-critical scenarios. But with Varjo’s human-eye resolution VR devices, the quality of the simulation now allows astronauts to see the necessary details clearly. Thanks to this technology, the Boeing Starliner program allows astronauts to train for all scenarios—from launch to docking to landing—completely in VR.

Varjo differentiates itself from other VR companies by their commitment to a new level of visual fidelity. Instead of a completely fabricated VR, they have also created a mixed reality in which VR objects are superimposed on top of the physical world, with the real and digital spheres merging seamlessly. “They need to be on parity, otherwise they don’t look real,” says Jussi Mäkinen, Varjo’s chief brand officer. “This is what drives our innovation forward.”

With Varjo’s video pass-through based mixed reality, virtual objects can look as real as real-world objects and reflect realistic lights and shadows. This image is an unedited shot-through capture of Varjo XR-3, portraying a virtual car superimposed on a real-world room.

One of Varjo’s goals was to make their technology more widely accessible by transitioning from physical computers to the cloud. Their mission was to create the Varjo Reality Cloud platform, which would allow real-time streaming to supported devices. In order to realize this vision, they needed a partner who was well-versed when it came to dealing with demanding customers and scaling quickly. With Amazon Web Services (AWS), they found speed, experience, and what they’ve described as an “extraordinary” closeness and support.

Jussi Mäkinen, Varjo’s chief brand officer.

“If you’re really serious about the metaverse, you need to build your own hardware, software, and service on top,” says Jussi. “From the very beginning, it was clear that we needed a partner who very deeply understands cloud technology and has dealt with the most demanding customers in this space.”

Jussi envisions the Varjo Reality Cloud ultimately doing for VR what Spotify did for music: creating a high-fidelity cloud streaming experience that will be accessible to anyone, anywhere.

Another key invention that has enabled seamless usage of VR and XR is what Varjo calls the “foveated transport algorithm.” This eye tracking technology uses a combination of optical innovation, mechanical innovation, and software and algorithm innovation, allowing data to move back and forth between the user and the cloud based on the user’s eye data. “We understand where you’re looking,” says Jussi. “We can actually predict that, so we can predict the pixels that you need to shift back and forth through internet connection.”

Going forward, the Varjo team is focused on creating the next computing paradigm using streaming services and wearable technology. “The world will change quite a lot,” predicts Jussi. “When things around you don’t need to have physical user interfaces, they can be virtual, and you don’t need to create so many physical things anymore.” Jussi hopes this will lead to a more sustainable future with less consumerism, not as much travel, and less waste. “I firmly believe that the creation of physical things will look very, very different in the future through this programmable layer that the world will have.”

The Varjo team also hopes to continue breaking down the boundaries between the physical and virtual worlds. Jussi says this requires a type of experimentation and creative thinking that would not be possible without AWS’s ability to scale and willingness to venture into unknown waters.

“I think AWS is one of the global icons in creating a new kind of marketplace,” says Jussi. “We really feel that we’ve found our soulmate for our journey of building and scaling a photoreal metaverse.”

Accelerating AI/ML scaling and AI development with Anyscale and AWS

Rob Ferguson — Thu, 17 Nov 2022 19:47:32 +0000

Building a cloud-distributed and scalable artificial intelligence (AI) application is a cross-team effort that requires complicated management of resources and comes with numerous production concerns such as code changes, refactoring, setting up the infrastructure, and complex developer operations (DevOps). These can confuse the development process, slow down time-to-market, and keep developers from focusing on product innovation.

“With the lack of established development-to-production pathways, some estimates place the AI development project failure rate at nearly 50%,” explains Rob Ferguson, Amazon Web Services (AWS) global head of AI and machine learning (ML) for startups.

That’s where Anyscale, Ray (the first unified, distributed compute framework for scaling ML or Python workloads), and the Anyscale Platform (a fully-managed Ray platform provided by the creators of Ray) help AI teams speed development and experimentation, and effortlessly scale applications to ensure the success of their AI projects.

Ray was conceived at the University of California Berkeley in the RISELab and open-sourced in 2017. Today, Ray is widely adopted by organizations globally for developing AI and Python applications at scale.

Improve scalability and developer ergonomics with the Anyscale Platform

To advance AI development and scaling for every organization, the Anyscale Platform extends the capabilities of Ray by enabling developers and cross-functional teams to accelerate experimentation and speed up the development of ML applications at scale. The Anyscale Platform and Ray increase developer velocity by providing scalable compute for data ingestion and preprocessing, ML training, hyperparameter tuning, model serving, and more, all while integrating seamlessly with the rest of the ML ecosystem.

The Anyscale Platform significantly extends the capabilities of Ray. Anyscale Workspaces provides developers with a unified and seamless development experience to scale ML workloads from a laptop to the cloud with no code changes. Developers can now leverage a single environment to build, test, and deploy workloads to production while leveraging the tools they are familiar with. The Anyscale Platform improves iteration speed by reducing cluster setup time by 5X over Ray. Finally, the Anyscale Platform provides observability, monitoring, and job scheduling out of the box.

“The capabilities of the Anyscale Platform go far beyond the capabilities of Ray and make AI/ML and Python workload development, experimentation, and scaling even easier,” said Ion Stoica, co-founder and executive chairman of Anyscale. “Thousands of organizations rely on Ray to scale their ML and Python applications, and a growing number of organizations are leveraging the Anyscale Platform to build, test, and deploy these applications to production faster than ever.”

From AI developers, to ML practitioners, to data scientists and engineers; all contributors can benefit from the ease, scalability, and developer ergonomics that the Anyscale Platform provides for the full AI journey: from model development, tuning, and training to inference and scalable model serving.

From left to right: Robert Nishihara, chief executive officer (CEO); Philip Moritz, chief technology officer (CTO); and Ion Stoica, executive chairman

Anyscale’s mission is to ease the building and scaling of AI/ML and Python workloads, as well as AI applications. Their team democratizes AI by making it easier for data scientists and ML engineers to work with project structures that make sense for building production ML systems. Scalable, unified, and open distributed computing, such as what Anyscale offers for Ray, helps lead to the mass democratization and industrialization of AI.

Anyscale’s commitment to open source and the freedom of shared technology is evident: The Anyscale team leads the Ray open source community by managing operations, reviewing contributions, and growing and developing the community. Additionally, they are the biggest contributor to Ray open-source software (OSS) and recently released Ray 2.1.

Building mature, reliable, and easily scalable ML models

With 90% of new applications expected to be cloud-native by 2025, Anyscale envisions the cloud as the default substrate in which the next generation of applications will be developed. They expect the next generation of cloud infrastructure to be optimized frameworks which ease the development and management of applications.

Anyscale leverages AWS’ mature suite of products as components for rapid development of their platform. Anyscale and Ray are full-featured on AWS, with highly efficient performance and scaling. In addition to AWS’ scale being a tremendous asset to Anyscale and Ray, some of their largest customers have a long history of using AWS and this facilitates their customer interactions.

“Inference is 90% of any ML project. Machine learning teams need novel ways of working with distributed computing to deliver the next generation of ML at scale,” explains Rob. “The golden era of machine learning is taking place and AWS and Anyscale help companies to build ML models the way they’re supposed to be: mature, reliable, and easily scalable.”

Amazon.com and AWS have worked with the Ray community to integrate Ray with many AWS services, including:

AWS Identity and Access Management (IAM) for fine-grained access control
AWS Certificate Manager (ACM) for SSL/TLS in-transit encryption
AWS Key Management Service (AWS KMS) for at-rest encryption
Amazon Simple Storage Service (Amazon S3) for object storage
Amazon Elastic File System (Amazon EFS) for distributed-file access
Amazon CloudWatch for observability

These service integrations allow AWS customers to scale their Ray-based workloads with secure, cost-efficient, and enterprise-ready AWS services across the AI and ML pipeline.

Solutions that fit within the AWS Well-Architected Machine Learning Lifecycle–such as the Anyscale Platform for the deployment stage—can help companies to spend more time on product development rather than resource provisioning and management.

The AWS well-architected ML lifecycle

The AWS Well-Architected Framework, which provides operational and architectural best practices for designing and operating workloads in the cloud, includes pillars for operational excellence, security, reliability, performance efficiency, cost optimization, and sustainability. With the Anyscale platform, developers can increase their performance efficiency.

“Anyscale simplifies a typical Python ML code and helps it fit within the framework within as little as two lines of code,” explains Rob. “AWS is an incredible home for Anyscale with its industry-leading tools for production systems as well as broadest and deepest compute options.”

Call for applications: AWS launches new startup program to support growth and innovation of French public sector startups

Asinetta Serban — Thu, 17 Nov 2022 17:14:51 +0000

Startups are key to accelerating the digital transformation of Europe’s public sector. French citizens are adopting digital government services at an increasing rate. According to a report by Public First, 82% of French citizens currently use e-government services, and 72% of French consumers said they were confident they could find information or a service through their government’s websites or apps.

Yet, while many startups develop new products to continue solving pressing societal challenges, instruments and mechanisms to support them are few. While initiatives (Citizens’ Initiatives Accelerator, Gouv.Tech Catalogue, APProch’ platform, etc.) have been launched in France to support emerging public sector startups, in 2021, according to the French Procurement Observatory, only 231 public procurements of innovative products and services have been concluded.

Public sector entrepreneurs are still facing hurdles with getting public sector contracts and access to related investment opportunities. In honor of the 9^th edition of the “Public Innovation Month” this November 2022, a series of events focused on sharing methods, best practices, and innovative solutions to design more effective public policies, headed by the French Interministerial Direction for Public Transformation (DITP), Amazon Web Services (AWS) is doubling its commitment to enhancing innovative ideas to support public sector transformation.

AWS is launching AWS Startup Ramp in France to accelerate the development of early stage startups in the public sector. This public sector tailored program will support startups in developing new products and services in government technology, healthcare, sustainability, smart cities, space technology, and more. AWS Startup Ramp provides technical mentoring, business development, marketing, networking support, and access to public sector-specific content to help the startups better understand public sector processes, such as preparing lengthy custom proposals and meeting intensive security requirements with vertical-specific certifications.

Initially launched in 2021 in India, Startup Ramp is also available to support startups in South Korea, Singapore, Indonesia, Malaysia, Thailand, Vietnam, the Philippines, and Japan. In Singapore, AWS Startup Ramp has supported NuSpace, a space tech company, navigate complex regulatory and security requirements, which has accelerated its development. Med247, a health tech startup in Vietnam, secured a $4.5M Series A funding round with the support of AWS Startup Ramp.

“Public Sector entrepreneurs, who have at heart a common interest, are building innovative solutions to help central, local governments, and citizens. They continuously inspire us to continue innovating along side them,” says Cameron Brooks, Europe leader for AWS worldwide public sector. “We are pleased that our AWS Startup Ramp will enable early-stage startups to turn their visions for change into disruptive technology to benefit public services. The program gives the great visionaries of today a path to becoming the future business brains of tomorrow – one we are excited to join them on.”

Katrin Dimitrova, co-founder of Vizzia, a French startup using artificial intelligence (AI) and video recognition to help local governments improve the way they manage waste, shared: “I’m excited to see AWS Startup Ramp being launched in France! We have received important support from AWS to grow our startup. We’re happy to see that it will benefit other startups like ours in the public sector sphere to continue building and offering essential services to communities, local administrations, and ultimately citizens.”

Roxanne Varza, director of Station F, the largest startup campus in the world, said “We don’t see enough startups trying to address the needs of the public sector – which could radically improve the lives of citizens. So happy to see this programme, which provides the additional support public sector innovation needs in France.”

Joining the program

How do I qualify?

AWS Startup Ramp will support startups in their earliest stages of development. To qualify, startups have to be founded within the last 10 years, have less than $10M in annual revenue, and plan to serve public sector customers. Startups building solutions in civic and government technology, smart cities, healthcare, sustainability, security, and space technology, are encouraged to apply to the program.

How do I apply?

Startup teams who are interested in participating in the first AWS Startup Ramp cohort with entrepreneurs across France will need to complete the following steps before applications close on 28 February 2023:

Submit the AWS Startup Ramp Application Request online.
Complete the AWS Startup Ramp application.
Indicate on the AWS Startup Ramp application that you are interested in being a part of the AWS Startup Ramp France cohort by marking the cohort checkbox on the main application.

The first AWS Startup Ramp France cohort will commence virtually later in 2023 with technical and business training sessions. The cohort will have the opportunity to engage with other entrepreneurs and their teams while learning from domain experts and AWS executives who are former founders themselves.

Request an application and apply to be an AWS Startup Ramp member today. You can learn more about the program benefits here.

Follow #StartupRamp on Twitter to connect with the community. If you have any questions, please contact the AWS Startup Ramp team directly.

Meet Astro — Astronomer’s managed Apache Airflow service built and hosted on AWS

Paola Peraza Calderon — Wed, 16 Nov 2022 13:04:58 +0000

For data to be useful in a modern enterprise, it must be collected and centralized from various sources, processed across a growing ecosystem of tools, and fed to systems across an organization in a way that’s consumable across teams. This data orchestration —weaving business logic through the data stack for everything from dashboards to personalization algorithms — requires hundreds, if not thousands, of data pipelines.

Data orchestration is needed across all industries, in organizations of all sizes. With more than 2,200 contributors and over 12M monthly downloads, Apache Airflow has emerged as the open source standard for programmatically authoring, scheduling, and monitoring data pipelines. Data practitioners love Airflow because of its community, its flexibility, and its ability to provide a central view of a data ecosystem.

However, data teams naturally need more than open source Airflow on its own — they need test pipelines to ensure data quality, SDKs to make data practitioners productive, and observability plus lineage for the underlying data — even as they strive to minimize operational overhead. Data lineage provides the full context of the data by capturing in greater detail the relationships between data sources, where the data originated, and how it gets transformed and converged through the data lifecycle.

Meeting the need for modern data orchestration

Astronomer, a startup founded in 2018, has spent the last five years advancing Airflow as an open source project with tools that help data practitioners get the most out of data orchestration and data lineage. Astronomer’s flagship product, Astro, enables customers to build, run, and observe data pipelines on Airflow as a managed service, which allows data teams to spend more time focusing on writing business logic and expanding access to data.

“Many fundamental business processes that Astro orchestrates for our customers are powered by Amazon Web Services (AWS): Amazon Simple Storage Service (Amazon S3), Amazon Redshift, Amazon EMR, Amazon SageMaker, and so many others,” explains Viraj Parekh, Astronomer’s Field CTO.

Co-founded by a small team that included three friends — Paola Peraza Calderon, Pete DeJoy, and Viraj Parekh — Astronomer describes its current mission as three-fold:

Build products that increase the value that data teams get from data orchestration and data lineage.
Cultivate the organic growth of the Airflow open source project and its community.
Provide education, best practices, and support to data practitioners that enable their success with data.

Pete DeJoy, product manager Viraj Parekh, field CTO Paola Peraza Calderon, product

manager

With more than 350 employees and a globally distributed team, both Astronomer and its customer base have grown quickly. “It started with people running open source Airflow and asking us for help with managing the infrastructure behind that,” Pete says. “Now that we’ve solved infrastructure management, we’re focused on the broader set of capabilities needed to take Airflow and use it as the foundation for a complete orchestration platform.”

Building and scaling on AWS

The market’s need for Astronomer products, as well as the company’s potential for success, was evident early on. Viraj laughs as he shares a story about their early days. “We were all hands on deck for a proof-of-concept with a large gaming company. The company relied on Astronomer to orchestrate the flow of data for its biggest launch of the year. The morning after the launch, there were no support tickets,” says Viraj. “And I thought, ‘Oh no, did something go wrong?’ Turns out, something went right. Everything worked. We were handling 100% of the data ingest that was coming from one of this company’s biggest launches, and everything ran without a hitch.”

Why did Astronomer build its startup on AWS? “I can’t say it was a decision. It was the obvious choice—AWS has been the cornerstone of our cloud strategy,” says Paola. “As a baseline, the ubiquity of AWS services across countries and regions allows us to work with organizations around the world. It single-handedly unlocks our market.”

To meet the broadening needs of its customer base, Astronomer builds interfaces that allow data practitioners to get the most out of Airflow as they develop data pipelines and form a singular view of their ecosystem. Viraj explains: “We’re merging data orchestration through any system you want—using whatever tools and services your team uses—with data lineage. Not only can you orchestrate data across all your systems, but you can see how that data moves.”

As shown in the architecture diagram, Astro is built with a multi-plane architecture that consists of a control plane hosted by Astronomer and a data plane that can run in your cloud or in a single-tenant account hosted by Astronomer:

The Astro architecture diagram

As Astronomer grows, the company has scaled its AWS footprint to meet the needs of its customers. Today, Astronomer relies on Amazon Elastic Kubernetes Service (Amazon EKS)to run Astro as a managed service within a customer’s corporate network, and supports tools like AWS Transit Gateway and AWS PrivateLink to securely connect to other data services in their network. Astro uses AWS CloudFormation to provision new Kubernetes clusters and Amazon S3 to store logs, and makes node instance types available for customers to choose the most optimal hardware to run their pipelines. This gives data practitioners optionality, performance, and efficiency where they need it.

“We’re confident that as our market and customer base grows, AWS can grow with us. Being able to fine-tune AWS services to fit our needs helps us make Astro faster, more cost effective, and easier to run for our customers,” says Paola.

Building a successful startup

For startups looking to replicate their success, the Astronomer founding team agrees that it’s critical to spend time with early adopters of the product. This creates a tight feedback loop that improves your product early on, and often results in strong personal relationships that will guide you throughout the company-building journey.

“Especially for early-stage startups, the people who adopt your product first are most likely to understand the problem you’re trying to solve. Curate those relationships over time, because these customers have been thinking about your problem and using your solution as long as you have.” – Viraj Parekh

“Ask a lot of questions — and put in the work. So much about taking a company through its early stages is about rolling up your sleeves, letting yourself iterate, and rallying a small team alongside you. As simple as it sounds, execution ultimately differentiates so many successful ventures.” – Paola Peraza Calderon

“As your company grows, the list of things to do will never end. It’s a real skill to learn how to identify what the high-priority items are on the list and focus on accomplishing those.” – Pete DeJoy

What’s next for Astronomer?

As for what’s next for Astronomer, Pete explains: “We want to build a generational company that creates real customer value, while cultivating talent among our employees, and allowing them to self-actualize in their careers. And we’re going to get there by driving tangible, meaningful customer outcomes on a day-to-day basis.”

How LabVoice + AWS are expanding accessibility in research labs

Bonnie McClure — Tue, 15 Nov 2022 18:23:06 +0000

As a PhD student in biophysics at Yale, Sara Siwiecki spends a lot of time in the lab working on the small details, like peering through a microscope or managing inventories of the chemicals she needs for experiments. But she often encounters obstacles due to outdated lab and equipment designs that don’t accommodate her visual disability. Opaque containers that make it difficult to read measurements, general lab-wide disorganization, and tiny text are just a few of the challenges she confronts during a typical day.

At Brown, PhD student Gabriel Monteiro da Silva has spent hours tracking down substances in the lab. It’s routine but crucial work: lab chemicals can be expensive, and if they hang around past their expiration date, they can become safety risks. Keeping track of dozens of compounds is made especially stressful by lab design that doesn’t account for Gabriel’s attention-deficit/hyperactivity disorder (ADHD).

That’s why they decided to collaborate with LabVoice—a digital lab assistant platform designed specifically for the research lab. Working together with the LabVoice team, they developed an inventory search solution that allows users to record information, like chemical location and amount, and then retrieve it later entirely through verbal prompts. Instead of hunting for test tubes and scouring drawers, scientists can ask LabVoice about a substance and instantly get the info they need. This can relieve major stressors related to a lack of accommodations for mobility, memory, sight, or other types of disabilities.

Streamlining lab work for scientists to do their best work

“I thought that LabVoice was just super cool as a concept,” says Sara. “It’s something that’s very useful and was not necessarily designed in the context of accessibility, but is super applicable to accessibility—which is kind of the most ideal situation that most people with disabilities want: something that is useful for everyone and is also applicable and useful to people with disabilities.”

Lab research also often requires significant amounts of repetitive, non-scientific work. Some studies suggest that up to 50% of researchers’ time is spent on administrative tasks. That can include things like preparing data for ingestion and preparing reports on work already completed—all tasks that can be streamlined with LabVoice.

More importantly, though, is the lost time and frustrations that accrue thanks to the reality of inaccessible lab environments. Due to various factors (like lack of funding), many labs are not equipped with up-to-date accommodations and fail to account for the diversity of disability experiences.

“There are labs that have existed for 60-70 years without any of the additions to accessibility that we have seen in shopping malls, doctor’s offices, buildings,” says Gabriel. “Because academia is so insulated and the people doing these kinds of upgrades—they either don’t consider it a priority, or even if they do, they usually are not properly funded for that.”

This requires scientists to work regularly in spaces that aren’t designed for their needs. By bringing a comprehensive AI voice assistant into these environments, LabVoice makes significant strides for accessibility in the research sciences. It empowers scientists to focus on their research rather than on accessibility gaps, routine chores, and administrative duties.

Amazon Lex and Amazon Polly help deliver a complete digital lab assistant experience

LabVoice is available as both a smart device and a mobile app. The digital assistant functions similarly to a smart speaker—think Amazon Echo for scientists—and can facilitate note-taking and data throughout the lab. It also offers features like the option to take photos, record videos, and scan barcodes, in addition to guiding a user through scientific processes (standard operating procedures (SOPs), protocols, checklists, etc). LabVoice links with existing lab software and equipment, allowing scientists to use instruments and record data hands-free. Its ease of use and connectivity with lab infrastructure can enhance regulatory and safety protocol compliance. And it’s a boon to accessibility.

All of this is made possible thanks to LabVoice’s integration with AWS. According to Steve McCoy, Head of Sales at LabVoice, LabVoice’s implementation of AWS services goes beyond cloud storage and computing—the AWS platform is fundamental to LabVoice’s operations.

In particular, LabVoice relies on conversational AI service Amazon Lex and text-to-speech software Amazon Polly to deliver a complete digital lab assistant experience. Steve explains that the team uses these tools as a foundation on which they develop additions and extensions that are science specific.

AWS’s expansive dataset also allows the team to continually refine the LabVoice experience. “By using AWS as the backbone of what we’re doing, we’re able to access pools of millions and millions of [users of] these products,” Steve says. This gives the LabVoice team a massive sample size of voices and vocal sounds to pull from as it works to improve the platform’s voice recognition capabilities. That data is critical to improving LabVoice’s ability to identify terms that are encountered regularly in the lab but may be uncommon elsewhere, like chemical compounds or scientific devices.

Steve believes that the future for LabVoice + AWS is a seamless experience that captures the entirety of a scientist’s stream of consciousness as they work. As the tech continues to develop, a researcher could “speak in real time with their observations, their notes, their measurements, and then use what AWS offers to just turn that into something that is automatically entered into their electronic lab notebook.”

These developments would be impactful for the day-to-day work of many researchers in labs that lack proper accommodations. “Obviously, academia is what millions of people do just in the US alone,” says Gabriel, so accessibility issues in the lab affect far more than “just a niche of people.”

“Looking forward to the future in general for science, I think things could be significantly improved and make everyone’s lives easier by implementing things like LabVoice,” says Sara. “There’s just endless amounts of opportunities that could come out of it.”

The technology could assist workers in a wide array of fields, from agriculture to marine biology. Whatever the industry, LabVoice will be an asset in allowing people like Sara, Gabriel, and their peers to continue working on the solutions of the future.

Advice from a unicorn founder – How to grow, sustain, and scale your startup

Ina Stuve — Mon, 14 Nov 2022 17:42:12 +0000

No matter what market you’re in, successful startups all have a few things in common—a passion and commitment for what they’re doing, a great story to tell and a laser-like focus on customer needs.

Dataiku has taken the data and AI world by storm—transforming from French startup to global unicorn in just seven years. The team’s journey began with a passion for data and machine learning (ML) and a quest to bring everyday artificial intelligence (AI) to companies of all sizes and sectors.

As a founder, your path will of course be different from Dataiku’s, but they can show you what to look out for and provide advice to aid and speed your progress.

Uncertainty is a fact of life—make it work for you

“When you’re on an entrepreneurial journey, you have to learn to be comfortable with uncertainty, because you can’t control and predict everything,” Dataiku co-founder and CEO Florian Douetteau points out.

He explains that you might not be able to do anything about the unpredictability and uncertainty of the world, but you can be proactive and take bold steps.

In the early startup phase, the Dataiku team chose to multiply the number of countries and time zones they operated in, which meant they spread the risk of failure.

Although the past couple of years have seen a huge appreciation of the value of data and AI across almost every industry and developed country, when Dataiku launched in 2013, that was not the case.

By deciding to expand outside their original base in France into key economies like the US and UK where digital-first business and e-commerce were more advanced, the company was able to grow faster.

Now that “everyday AI” is more established around the world, Dataiku is in a stronger position, because in so many countries and across time zones.

Their global reach also provides insurance against the uncertainty of the global economy. If one country’s economy is struggling, Dataiku can focus on other countries that are enjoying stronger growth.

Dataiku’s gamble to spread across time zones paid off. They’ve built a business with more than 1,000 employees spread across 15 countries, including the US, the UK, Germany, the Netherlands, Australia, and Singapore.

Driving growth globally means riding the wave

By keeping up to date with trends and going with the market’s flow, you can ensure your company has a long-term purpose and viability. Think of it like this. When you launch a rocket, it takes off towards its destination, but if it encounters obstacles or interference, it must continually correct its trajectory.

The past few years have seen a massive acceleration of the use of data and AI within enterprises of all types and sizes. Embracing this technology has helped Dataiku to stay on course.

Florian points out that 10 years ago, “some people were doubtful about data,” which required the Dataiku team to hold their nerve and keep believing they were on the right track.

Ultimately, he says their “north star” was their customers. They helped guide Dataiku’s journey by giving feedback about what type of data they needed. Having that close relationship helped them stay on course and keep developing products that were useful and practical.

“Keep your focus, trust your customer and make sure customers trust you, so you can follow the trend,” Florian advises.

Create an authentic brand story early in your startup journey

Formulate the story of your company and the transformation you’re making and do it early in your journey.

Instead of just spouting facts and figures, make your message memorable by weaving it into a brand narrative that shares your setbacks, challenges, successes, and value proposition. As humans, we’re programmed to love stories. Neuroscientists have learned that stories trigger the release of the natural feel-good neurochemical oxytocin. Compelling stories can inspire brand loyalty as they help customers to connect with your company.

It’s a lesson Florian has learned along the way: “When we started the business, we were a bit shy and not so sure how to tell this story about business users going into AI” he says. I wish I’d known earlier how important it is to tell and build the story of your business.”

That said, Florian warns it isn’t just about coming up with buzzwords. H “intellectually dubious about a situation where you reduce culture to a few words.”

For example, he explains “as soon as you describe yourself as ‘humble’, you’re not humble at all, so it’s not always smart to claim something like that.”

Instead, he recommends the team sitting around the table to have a series of genuine discussions about the culture, while continuing to build your own narrative by acting in “an original and unconventional way.”

If you want to impress investors, differentiate yourself

One of the most important aspects of entrepreneurship is to focus on your differentiating factors.

In Dataiku’s case, it was building an end-to-end platform that can be leveraged by almost everyone in an organization—not just data scientists.

“It’s important to have this distinctive position, because it’s the way you convince investors to finance your project,” Florian says. “If you’re not clear about what sets you apart, you’ll be lost in the sea of competitors who will all seem to be doing the same thing as you.”

But when you are deciding what to make your speciality, remember you’re at the beginning of a long journey, and your idea must be something you’re passionate about.

As Florian puts it: “Let’s imagine for a second that your startup is successful—you’ll need to be ready to work on that specific topic for 15 years. There’s no point in trying to fake it, because it’ll come back and bite you.”

Scale up faster by harnessing the power of the cloud

Most organizations have to manipulate huge amounts of data. Typically, the most efficient way to do this is to run it in the cloud. With Amazon Web Services (AWS) tools, you can scale up and down when and where you need extra capacity, and run analytics rapidly.

From its earliest beginnings, the Dataiku team built its product on top of AWS services. They built it to focus on the need of business users, concentrating on ease-of-use and user interface. Particularly, they focused on a broader user base in enterprises who don’t typically have developer skills.

“We didn’t focus as much on infrastructure, storage, and data. You could say we built Dataiku as a ‘lazy product’, because we push the computation to the cloud, “ Florian explains. “It’s a UI, where you click around. It’s a very interesting and smart UI but ultimately, as much as possible, we push it to the cloud.”

AWS services like AWS database services storage, compute, the Amazon Elastic Kubernetes Service (Amazon EKS), and Amazon Redshift helped Florian and the team set up Dataiku. “Without the cloud, it would not have been possible to focus on what we wanted to do,” he adds.

Dataiku’s partnership with AWS has many benefits, as Florian points out: “All our customers benefit from the cloud because the cloud is what makes our business possible.

“Our customers have to manipulate terabytes, or even petabytes of data to build analytics or manipulate the data.” Because these analytics need to happen at large scale, Florian believes the best way to make that possible and efficient is to run it in the cloud. This way, Dataiku and its customers benefit from the ability to scale up and scale down and run analytics computation rapidly.

Operating through gives Dataiku agility. “You need the agility of the cloud to innovate, because you can’t predict everything in terms of success, cost, or even the amount of data or analytics to be delivered and used during the first iteration of those projects,” Florian adds.

Don’t measure productivity by the number of meetings

Carving out dedicated time to think or read is crucial, not just for your sanity but because it’s how you’ll find true inspiration.

It’s easy to find yourself in back-to-back meetings, especially when you’re operating in more than ten time zones. Unfortunately, when that happens, there’s often no time to come up with great insights.

Finding time to relax is key. As Florian puts it: “What I’ve learned is that finding time to relax, and sometimes just stare out of the window, is what makes me truly productive.”

Put your customers first in everything

Really focus on the quality of your product and service and the way your customers perceive it. In other words, take time to make sure your customers perceive your product as being of value to them and that they are happy.

Commit to making your customers feel valued, understanding their needs and how they’re using your product or service, being responsive at all times and making sure you solve any problems they have quickly.

This link between product and customer needs to be deeply embedded in your company’s culture and something you hang onto, no matter how big your enterprise grows.

Quickly deploy and skillfully manage your databases with the AWS Database Plug & Play Program

John Nguyen — Fri, 11 Nov 2022 19:53:47 +0000

It’s fairly typical for startups not to have dedicated resources to manage databases – in fact, you may not even have a database yet! That’s totally fine. Databases are complex systems that require expertise in a range of disciplines, including hardware, networking, storage management techniques, capacity planning, etc. In an enterprise setting, you would likely have a dedicated person or team (database administrator) to create and maintain databases.

So, what do you do if you don’t have the resources, time, or funds to self-manage a database? Use Amazon Relational Database Service (RDS)! Amazon RDS allows you to set up, operate, and scale a relational database in the cloud with just a few clicks. It removes inefficient and time-consuming database administrative tasks without needing to provision infrastructure or maintain software. And, with the new AWS Database Plug & Play Program, you’ll get a packaged bundle of AWS advisory time, architecture and prototyping patterns, AWS usage credits, and pathways to go-to-market acceleration programs to help further ensure your teams can focus on value-generating work.

Validated CloudFormation templates for quick deployment

The program includes AWS CloudFormation templates, available on GitHub, that create the following open-source Amazon RDS database (Figure 1):

Figure 1.

We’ll explain these templates and their benefits in detail in the next section, but for now, here’s a summary of the Cloud Formation templates and what they offer.

The CloudFormation templates provision the network infrastructure and all the components shown in Figure 1. The CloudFormation templates are split into three stacks.

Template to set up Amazon Virtual Private Cloud (VPC), subnets, route tables, internet gateway, NAT gateway, Amazon Simple Storage Service (Amazon S3) gateway endpoint, AWS Secrets Manager interface endpoint, and other networking components.
Template to set up a Linux Bastion Host in an Amazon Elastic Compute Cloud (Amazon EC2) Auto Scaling group to connect to the RDS DB cluster.
Template to set up any of the following databases: Amazon Aurora Serverless v2, Amazon Aurora PostgreSQL-Compatible Edition, Amazon RDS for PostgreSQL, Aurora MySQL-Compatible Edition, Amazon RDS for MySQL, with master user password stored in Secrets Manager and bootstrap the database using AWS Lambda.

Using three different CloudFormation stacks instead of one nested stack gives you flexibility. For example, you can choose to deploy the VPC and bastion host CloudFormation stacks once and Aurora PostgreSQL DB cluster CloudFormation stack multiple times in an AWS Region.

The README of the GitHub repository has detailed template usage instructions. Currently, it contains instructions for the following engines: Amazon RDS for PostgreSQL, Amazon RDS for MySQL, Amazon Aurora PostgreSQL, Amazon Aurora MySQL, and Amazon Aurora Serverless v2 (PostgreSQL/MySQL). Future updates will include: Amazon DynamoDB, Amazon DocumentDB, and Amazon Neptune.

What benefits will I get from this program?

By enrolling in the program, you will receive the following benefits:

Validated architecture and prototyping patterns. Take advantage of the pre-built CloudFormation templates for quick deployment and rapid prototyping.
Curated self-service training. The training is available at https://skillbuilder.aws.
AWS Solutions Architecture Advisory. Receive up to two hours of consulting time with an AWS Database Specialist Solutions Architect on a range of topics, including architecture, migration, and prototyping.
AWS credits for prototyping. Receive up to $2,000 of AWS credits to run a proof-of-concept using AWS RDS.
Pathways to go-to-market. Access to AWS Software Partner Path for co-market and co-sell program support.

Technical assets

The CloudFormation templates create Amazon RDS resources and have best practices included in each CloudFormation template to ensure you have an optimal architecture.

The VPC template takes care of the following:

Sets up three Availability Zones for high availability and disaster recovery. Availability Zones are geographically distributed within a Region and spaced for best insulation and stability in the event of a natural disaster.
Provisions one public subnet and one private subnet for each Availability Zone. We recommend using public subnets for external-facing resources and private subnets for internal resources to reduce the risk of exfiltration of data.
Creates and associates network ACLs with default rules to the private and public subnets. We recommend using network ACLs as firewalls to control inbound and outbound traffic at the subnet level. These network ACLs provide individual controls that you can customize as a second layer of defense.
Creates and associates independent routing tables for each private subnet, which you can configure as necessary to control the flow of traffic within and outside the Amazon VPC. The public subnets share a single routing table because they all use the same internet gateway as the sole route to communicate with the internet.
Generates a NAT gateway in each of the three public subnets for high availability. NAT gateways offer major advantages over NAT instances in terms of deployment, availability, and maintenance. They allow instances in a private subnet to connect to the internet or other AWS services while they prevent the internet from initiating a connection with those instances.
Produces an S3 VPC endpoint, which provides resources in private subnets. For example, it allows Lambda to communicate with Amazon S3 in a secure and reliable way.
Creates Secrets Manager interface VPC endpoint, which provides Lambda resources in private subnets to communicate securely with Secrets Manager without requiring internet access.

The bastion host template takes care of the following:

Creates an Amazon EC2 Auto Scaling group that’s spread across the three public subnets that were set up by the VPC CloudFormation template. The Amazon EC2 Auto Scaling group ensures that the bastion host is always available in one of the three Availability Zones.
Sets up an Elastic IP address and associates it with the bastion host. The Elastic IP address makes it easier for on-premises firewalls to remember and allow these IP addresses. If an instance is terminated and the Amazon EC2 Auto Scaling group launches a new instance in its place, the existing Elastic IP address is re-associated with the new instance. This ensures that the same trusted Elastic IP address is used at all times.
Sets up an Amazon EC2 security group and associates it with the bastion host. This allows you to lock down access to the bastion hosts to known Classless Inter-Domain Routing (CIDR) scopes and port for ingress.
Creates a CloudWatch Logs log group to hold the bastion host’s shell history logs and sets up a CloudWatch metric to track SSH (Secure Shell) command counts. This helps in security audits because it allows you to check who is accessing the bastion host.
Creates a CloudWatch alarm to monitor the CPU on the bastion host and send an Amazon Simple Notification Service (Amazon SNS) notification when the alarm is activated.

The Amazon RDS template takes care of the following:

Creates a Multi-AZ RDS DB with a primary instance and an RDS DB replica in two separate Availability Zones for a production or pre-production type of environment. We recommend using this for high availability because the RDS DB automatically fails over to an RDS DB Read Replica if the primary DB instance becomes unavailable.
Places the RDS DB in private subnets. To access the DB, use the bastion host, which is set up by the bastion host template.
Sets up an Amazon EC2 security group and associates it with the RDS DB. This allows you to lock down access to the DB to known CIDR scopes and port for ingress.
Generates a random admin user password using Secrets Manager and associates this password with the RDS DB. A Python-based Lambda function backed by CloudFormation custom resource configures automatic password rotation every 30 days using Secret Manager. We recommend rotating passwords regularly to prevent unauthorized access in case the password is compromised.
Creates a DB parameter group with suggested settings and associates it with the RDS DB instances. The DB parameter group settings are provided as general guidance and should be reviewed and customized to suit your needs.

Ready to get started?

The Database Plug & Play Program is intended to be used with AWS database technologies to help manage, maintain, and ultimately unlock the value of your data. Get started today!

Register at https://pages.awscloud.com/ContactFormforPlugnPlay.html to contact team to schedule a consulting session or check eligibility for AWS credits.
Download the Program assets from GitHub (requires a GitHub account). Assets include the CloudFormation templates described in this post, recommended best practices, and training links.
To learn more about the AWS Partner Network (APN) and go-to-market program support, self-register on APN. There is no cost to self-register.
For more information, please reach out to your AWS Account Manager or send an email to dbplugandplay@amazon.com.

How Amazon Textract helped Fyle boost data extraction accuracy

Madhav Mansuriya — Wed, 09 Nov 2022 22:50:17 +0000

Fyle is an intelligent spend management product that helps businesses keep track of their expenses. They eliminate all manual activity around expense management. They know it’s frustrating for employees to fill out expense reports with details like amount, category, etc., because most of these details are already available in the receipt. With their Data-Extractor service, when an employee uploads the receipt, fields like amount, currency, date of spend, category, vendor, etc., fill in automatically.

Originally, the Data-Extractor service relied on an external service provider for optical character recognition (OCR) and Fyle’s internal machine learning algorithm to detect amount, category, date, currency, and vendor information. Unfortunately, they were receiving some feedback from customers that their tool wasn’t very accurate. As you can imagine, this isn’t the best place to be, so they rewrote their Data-Extractor service to use Amazon Textract because of its intuitive web console for APIs, which allowed them to test APIs in real-time with personalized input. This let them quickly try out an Amazon Textract API, which helped them achieve their goal of turning around a solution in two months. After implementing their new solution, Fyle saw 51.7% improvement in accuracy for the Data-Extractor service.

Testing the original solution

Before embarking on any major changes, Fyle built a test suite to systematically measure the accuracy numbers of the Data-Extractor service. They initially built out the data using humans, which took about a month. After that, they compared the accuracy of the Data-Extractor service against the human dataset. Table 1 shows the results.

Table 1. Data-Extractor’s performance before using Amazon Textract
Receipt features	Extraction accuracy in %
Date	43.46
Amount	45.37
Category	16.53
Currency	46.04
Vendor	21.89

Admittedly, these numbers are not great. So, Fyle embarked on improving accuracy. They had two options: 1) make incremental improvements or 2) rewrite the service. Fyle was leaning towards a rewrite for maintainability reasons, but then they found Amazon Textract.

Applying Amazon Textract to improve accuracy

Fyle’s team of two engineers working on this project were tasked with producing a solution within 2 months. This seemed like a pretty difficult ask, especially since they needed to also support the current version of the Data-Extractor service that was in production.

But after finding Amazon Textract, they saw that they could quickly test APIs and that all of Amazon Textract’s API contracts and code examples in different languages are listed with a detailed explanation of possible errors. This made the task considerably easier.

Once they decided to use Amazon Textract, all Fyle had to do was to convert their files into the required format for particular Amazon Textract APIs, AnalyzeExpense and DetectDocumentText, and make a request to the APIs from their code.

Amazon Textract AnalyzeExpense API extracts financial information

The AnalyzeExpense API synchronously analyzes an input document for financially related relationships between text. This helps Fyle extract information from uploaded receipts, including:

Vendor Name: VENDOR_NAME
Total: TOTAL
Receiver Address: RECEIVER_ADDRESS
Invoice/Receipt Date: INVOICE_RECEIPT_DATE
Invoice/Receipt ID: INVOICE_RECEIPT_ID
Payment Terms: PAYMENT_TERMS
Subtotal: SUBTOTAL
Due Date: DUE_DATE
Tax: TAX
Invoice Tax Payer ID (SSN/ITIN or EIN): TAX_PAYER_ID
Item Name: ITEM_NAME
Item Price: PRICE
Item Quantity: QUANTITY

Amazon Textract extracts these fields with a confidence score to show users how accurate the data pull is. It also provides the location (geometry coordinates) of the text on the receipt, so that users can further use that information for tagging features on the receipt.

DetectDocumentText

The DetectDocumentText API detects text in the input document. This helps Fyle get detailed OCR of entire receipts. It provides OCR in the form of lines and words with the confidence score of how confident Textract is with the extraction.

Figure 1 shows a sample output.

Figure 1. Sample output of Amazon Textract DetectDocumentText

Fyle rebuilt their data extraction service on top of Amazon Textract. With this service, they were able to write a considerably more maintainable version of their Data-Extractor service that worked across paper and the digital receipts with automated tests and 87% code coverage.

After applying Amazon Textract, they compared the results against human accuracy and their Data-Extractor service, as shown in Table 2.

Table 2. Data-Extractor’s performance after using Textract
Receipt features	Extraction accuracy in %
Date	74.85
Amount	68.84
Category	44.47
Currency	72.18
Vendor	47.12

Figure 2 shows a visualization of the change in accuracy after the Amazon Textract integration.

Figure 2. Old vs new accuracy for receipts

Fyle rolled out the change to all their customers over a 10-day period without any major issues. Customers even noticed the difference and sent thank you notes on the improvements.

Conclusion

With Amazon Textract and good engineering practices, Fyle wrote a considerably more maintainable version of the Data-Extractor service with a small team. It took them a few steps further on their mission of removing the drudgery associated with expense management for their customers.

Four simple steps to classify your data and secure your startup

Neil DCruz — Mon, 07 Nov 2022 17:15:20 +0000

Understanding your data types and their sensitivity levels ensures that your startup stays ahead of unintended data use or disclosures and satisfies compliance requirements.

To get you started, this post provides four simple steps to simplify and automate the data classification process for your startup.

1. Design your data classification framework

First, you’ll establish who is allowed access to your startup’s data by assessing how to classify and control your data.

Assess and classify your data

Data classification “involves identifying the types of data that are being processed and stored in an information system owned or operated by an organization,” as defined by the Data Classification Amazon Web Services (AWS) whitepaper.

To design your data classification framework, you’ll assess your data first. To do this, we recommend grouping data into as few categories as practical, because a simpler framework is easier to manage.

For most companies, these three tiers should cover the most common data classification use cases:

Unclassified. For low-security data (such as public website data, course catalog, etc.)
Official. For moderate-security data (such as official communication, internal memos, etc.)
Secret. For the most sensitive data (such as financial records, intellectual property, legally privileged data, etc.)

How you classify your data and where your store it will be informed by the nature of your business and its legal requirements. For instance, if you process credit cards, you must comply with payment card industry (PCI) standards, which are classified as Secret.

Need some help classifying data? Amazon Macie uses machine learning (ML) to automate the classification of sensitive and business-critical data. It uses machine learning (ML) and pattern matching to detect sensitive information like credit card numbers, health data, and other kinds of personally identifiable information (PII).

Determine data protection controls

Data protection controls manage how your data is used, who has access to it, and how it is encrypted.

After you assign your datasets to the classification tiers, you’ll determine what controls apply to each category. By carefully managing an appropriate data classification system, along with each workload’s level of protection requirements, you can map the controls and level of access or protection appropriate for the data.

There are many types of data protection controls, but some common ones you need to know are:

Using separate accounts to place workload resources per sensitivity level. AWS Organizations allows you to create and manage multiple AWS accounts with ease.
Setting up IAM policies, Organizations service control policies (SCPs), AWS Key Management Service (AWS KMS), and AWS CloudHSM, allows you to define and implement your policies for data classification and protection with encryption.
Ensuring that your data is stored and processed in the appropriate AWS Region based on your compliance and data residency requirements. AWS never initiates the movement of data between Regions. Content placed in a Region will remain in that Region unless you explicitly enable a feature or use a service that provides that functionality.
Using AWS Config rules to check automatically that you are using encryption for data at rest, for example, for Amazon Elastic Block Store (Amazon EBS) volumes, Amazon Relational Database Service (Amazon RDS) instances, and Amazon Simple Storage Service (Amazon S3) buckets.
Enforcing encryption in transit. For instance, HTTP requests can also be automatically redirected to HTTPS in Amazon CloudFront or on an Application Load Balancer.

Unclassified data, for instance, might be available to anyone in your organization or even externally, while Secret data might require authorized access to a key in order to decrypt.

2. Tag your data

For security purposes, metadata tags can help you identify, categorize, and manage resources in different ways, such as by purpose, owner, environment, or other criteria. You can use security tags to identify and group your resources based on confidentiality and compliance requirements.

Now that you know how you’ll classify and protect your data, you’ll add metadata tags that designate its classification level. These tags provide a reference for your team and enable additional automation and controls. There are many common tagging strategies, but the tags you might want to use for security include:

Confidentiality –An identifier for the specific data confidentiality level a resource supports
Compliance –An identifier for workloads that must adhere to specific compliance requirements

You can use tags to require encryption, for example, or restrict who can access data. If you do not already have tags for your existing resources, AWS offers tools to help manage resource tags across multiple services:

AWS Resource Groups and the Resource Groups Tagging API enable programmatic control of tags, making it easier to manage, search, and filter tags and resources.
AWS Identity and Access Management (IAM) allows you to control access to resources using tags.

Remember: It’s fairly easy to change tags to accommodate changing business requirements, but consider the consequences of future changes. For example, changing access control tags means you must also update the policies that reference those tags and control access to your resources

3. Redact sensitive data

Data redaction limits the places where sensitive data is stored and restricts access to those that need it, without hindering downstream environments.

Automating redaction limits the danger of inadvertent data releases and helps startups comply with privacy requirements as they navigate a growing volume of data.

Amazon Comprehend, in conjunction with Amazon S3 Object Lambda access points, can detect more than a dozen types of PII, including passwords, addresses, phone numbers, and Social Security numbers and automatically make redactions in text documents.

4. Keep up with compliance

Data compliance identifies rules for data protection, security, and storage. It establishes policies, procedures, and protocols and ensures data is protected from unauthorized access and use.

In our final step, you’ll monitor your data to ensure it maintains compliance. AWS Config Rules and AWS Foundational Security Best Practices in AWS Security Hub help you keep track of how your data settings are configured and notify you if they’re changed. They enable you to continuously ensure your security standards are being met.

Securing your startup with AWS

Almost every startup now operates in a global environment, and increasingly, they face differing requirements in each country in which they do business.

By identifying the data you have and implementing appropriate, automated controls, you can meet these requirements more easily, while also improving your security posture. For further reading, we recommend the AWS Data Classification Whitepaper. We also recommend reviewing the AWS Well-Architected Framework, which helps you understand the pros and cons of the decisions you make when building systems in the cloud and describes how to take advantage of cloud technologies to protect data, systems, and assets in a way that can improve your security posture.

Scale your startup with Serverless on AWS

Matthew Meckes — Fri, 04 Nov 2022 15:24:01 +0000

In 2019, cinch launched as a consumer-facing online platform that helped customers in the UK find, buy, and sell used cars. In late 2020, cinch spotted a trend in the used car sales market with new players going direct-to-consumer in the US and Europe. The pandemic accelerated the trend, and customer demand for used cars. Because of this, cinch announced a shift of their business model towards a direct-to-consumer car retail marketplace. The opportunity was huge, but required a domain (and organizational) transformation.

To deliver their new vision, the cinch team built a prototype on their existing containers platform. But with fewer than 30 employees, they got bogged down by managing it. After experimenting with different options, they decided to pivot to Serverless on Amazon Web Services (AWS) framework to reduce complexity. “Once I saw that an application was one file in serverless framework against hundreds in Kubernetes. It was a no-brainer,” says Jaz Chana, Technology Director at cinch.

By using AWS to scale their infrastructure, the cinch team was able to focus their cognitive load on improving the platform, quickly release new features, and re-build existing ones based on real world customer insight.

With their new architecture, cinch was able to pivot their business to the new model in 6 months, increase traffic by 2.5x (6,000 to 16,000 requests per minute), and reduce latency. They went from hundreds of cars sold within days, and grew by a factor of 100x within a few weeks.

Save time and effort with a walking skeleton

To save time in the initial architecture and design, the team built a “walking skeleton”: a tiny implementation of a larger system that performs a small end-to-end function. A walking skeleton doesn’t aim for perfection, but must be robust enough to pull together its main architectural components. The architecture and the functionality can then evolve in parallel.

As Figure 1 shows, the walking skeleton initially relied on Amazon DynamoDB, Amazon Simple Storage Service (Amazon S3), and an AWS Lambda function. The Lambda function loaded an array of vehicles from Amazon S3 and performed any sorting and filtering in memory. This approach worked for simple filtering and supported a few thousand cars.

Figure 1. Architecture of the first version of the cinch search component

But as the application gained popularity and the team gained real world insights, they saw that the existing architecture couldn’t support the complex aggregation/facet counting on multiple selectable filters that customers wanted.

With serverless you can rebuild instead of refactor

As shown in Figure 2, to address the increased requirements, the team created a new repository instead of iterating on their existing codebase. They also introduced Amazon OpenSearch to deliver advanced filtering at scale. Amazon EventBridge decoupled the search service from the rest of the platform. This let the search team replace the service without affecting other teams.

Figure 2. Architecture of the second version of the cinch search component

With this revised architecture, cinch can autonomously maintain integrations across domain boundaries. And, as they continue to scale, the on-demand, usage-based pricing model of serverless services allows them to adopt or decommission services based on their requirements, without having to consider contract terms or the cost of running multiple architectures in parallel.

“It might sound like this approach involved more effort, but it ended up being much cleaner. Ultimately, I don’t even think it required more work. Starting from scratch without serverless would have taken much longer; we may not even have done it at all.” Bertie Blackman – Automation Engineer – cinch (Search Team)

Reducing cognitive load to save time

To maximize the output of the team, cinch aimed to optimize for cognitive load, or the amount of information that working memory can hold at one time.

There are 3 types of cognitive load:

Germane cognitive load is the effort that is core to the topic. By increasing germane load, you can solve the problems that matter.

For example, “What service should offer vehicle details?”

Intrinsic cognitive load is the effort associated with building a software platform. By decreasing intrinsic load, you’ll be able to scale the team faster and easier.

For example, “How do I consume an event? How do I add something to a queue? How do I test a Lambda function?”

Extraneous cognitive load is related to the environment in which the task is being done. You want to get rid of extraneous load through automation and by offloading everything your application needs to do but doesn’t increase its competitive advantage in the eyes of its customers.

For example, “How do I configure this service to scale with traffic? What operating system should I use?”

Serverless speeds up decision-making to optimize for germane load

Serverless frees up cognitive load and allows teams to focus on the business domain. To support this, cinch defined a path to maximize memory for germane cognitive load, as summarized here and shown in Figure 3:

Test-driven development: teams create automated test cases alongside code, eliminating the need for a separate test team
Observability: instrumenting the code with intent so that the teams can measure the health of business transactions
Pairing, mobbing: use agile software techniques that reduce bottlenecks handoffs and ensure high quality
Trunk-based: use source control branching model, with developers collaborating on code in a single branch
Serverless: build and run applications without managing infrastructure
Event-driven: build architectures with EventBridge to decouple services without having to manage implementation details

Figure 3. cinch’s path

Using Lambda, DynamoDB, and Amazon API Gateway encourages teams to build loosely coupled architectures and independently deployable services. With serverless, engineers don’t have to worry about managing AWS Availability Zones, scaling infrastructure, or operating system patching, which add extraneous cognitive load.

The cinch team uses EventBridge for inter-domain service communication. With EventBridge, when they have an event to share across domains, they can publish it to a shared event bus and not have to think about how it will be consumed.

Equally, when a team wants to respond to an event from another team, they know the event will be on a central shared bus. Having a shared location and event structure further reduces intrinsic cognitive load.

Figure 4. Types of cognitive load applied to cinch’s business

Conclusion

cinch achieved its ambitious goals by using AWS serverless technologies. When the new website was launched, traffic quickly increased by a factor of 2.5x (peak of 6,000 requests per minute to 16,000 requests per minute).

With their serverless architecture, AWS manages the scale automatically. So, as the number of requests increased, latency went down. Traffic is highly variable, with anywhere between 250k and 4.5 million events a day, a 15x difference in volume between quiet and busy days.

With AWS scaling the infrastructure, cinch can focus their cognitive load on improving the platform. Now they can quickly releasing new features and re-build existing ones based on real world customer insight. This helps fuel exponential growth; cinch has gone from hundreds of cars sold within days, to growing by a factor of 100x within a few weeks.

Related resources

Six reasons for startups to attend AWS re:Invent virtually

Alexa Korach — Wed, 02 Nov 2022 12:02:51 +0000

It’s almost the most wonderful time of the year again: AWS re:Invent is just around the corner. And while we’d love to hang out with you in Las Vegas from November 28 through December 2, never fear if you can’t attend in person.

You can still get everything re:Invent has to offer by attending virtually. (Well, except for the slot machines and general Vegas-fueled sensory overload, of course!)

Here are six reasons to attend Virtual re:Invent 2022.

1. Virtual re:Invent is 100% free

That’s right—you don’t even have to buy a ticket to watch leadership sessions, reap the wisdom and knowledge of our brilliant keynote speakers, and participate in breakout sessions.

2. You can participate from anywhere

Gaining virtual access to re:Invent means that you can participate from anywhere across the globe, no matter what time zone you happen to be in. You can also enjoy everything the conference has to offer while wearing your softest pair of pajamas, all from the comfort of your own home.

3. The entire re:Invent conference is recorded

re:Invent is recorded, which means that you can hit pause to put the kids to bed, walk the dog, or even take a meeting. Plus, you’ll still be able to enjoy access to leadership, keynote, and breakout sessions on-demand at any time after the conference, making it easy to revisit that talk you found particularly inspiring whenever you need a little boost.

4. More languages and closed captioning

Some of the measures we’re taking to make virtual re:Invent more friendly to all include closed-captioning in English, as well as translation (in French, German, Italian, Japanese, Korean, Mandarin, Portuguese, and Spanish) for all keynote and leadership sessions.

5. Nightly recaps and live coverage

Because we adore the members of our international community, we’re once again instituting AWS on Air, a nightly program that will recap highlights from the conference in non-English languages.

Similarly, theCube—which will offer virtual attendees live event coverage, focusing specifically on enterprise technology thought leadership—is back! You can find both AWS on Air and theCube on separate dedicated channels via our virtual event platform.

6. AWS DeepRacer racing league

If you participated in Virtual re:Invent last year, you probably enjoyed watching AWS DeepRacer, the world’s first global autonomous racing league. Well, it’s back, and the competition is even more cutthroat than last year! Virtual attendees can watch developers of different backgrounds and skill levels compete in the races for both prizes and fame, live or on-demand. We promise you won’t want to miss it!

See you there!

As always, Virtual re:Invent is going to be a blast, and now you have six more reasons to attend. Check out our landing page for a complete list of the 2022 sessions and other pertinent info. We can’t wait to see you there!

AWS EMEA Startup Loft Accelerator marks its one-year anniversary!

Kellen O'Connor — Wed, 02 Nov 2022 00:09:41 +0000

It has been an amazing first 12 months for our AWS Startup Loft Accelerator program. So far, we have helped over 275 early-stage (two years and under) companies with technical and business expertise, tailored training, and mentoring across Europe, the Middle East, and Africa (EMEA). Every month, we welcome 30 new startups into the 10-week virtual program. They learn together and form a supportive community of innovative entrepreneurs.

All the startups gain access to two training tracks, one focused on business and one on technology. They also get one-to-one advice from a pool of 60 mentors and can attend group sessions with other founders. The new companies can also draw on a rich library of video assets and other content tailored to their specific business needs.

We’ve had over 2,400 applications and nearly 80 percent of them are at bootstrap or early stage of fund raising. AWS has also built a network of over 100 venture capitalists within the program who can request introductions to specific companies.

One interesting theme we’ve seen from the hundreds of companies that have joined the program is the role artificial intelligence (AI) is having in startups looking to disrupt established industries like healthcare. For too long, AI has been the preserve of large companies with big budgets and research labs. But our cohort of young companies is showing that cloud-based solutions can bring the best of AI to even the smallest businesses.

Let’s hear more from the companies themselves.

Novu.ai is using AWS for more accurate x-ray analysis

Novu.ai is a German startup created from a university research project. It uses AI to help radiologists analyze X-ray images faster and more efficiently when looking for potential signs of breast cancer. With the number of radiologists across Europe falling, delays in results, burnout, and misdiagnosis are increasing problems, as highlighted by the British Medical Journal in the UK.

Anything that can improve assessment accuracy has a dramatic impact on health outcomes. Some 80 percent of women who go through the pain of having a false positive result will disengage from future screening programs, with potentially devastating impacts on their health. Novu.ai uses AI to rapidly categorize X-rays as “normal,” “suspicious,” or ”with findings” so that radiologists can focus their time in the right place.

Dr. Sherief Emam, co-founder and chief executive officer of Novu.ai, appreciated the way the AWS Startup Loft Accelerator program helped bridge the gap between their engineering and programming skills to bring a working solution to market. “Only one percent of medical AI research projects ever make it to market,” says Emam. “There are so many entry barriers that engineers and developers are not aware of. The program really helped us overcome these by thinking about what our customers really want, about software as a service, and other licensing options.”

Novu.ai engineers also benefited from technical sessions. “It was great to get instant access to the right people within AWS,” says Emam. “The program gave us amazing one-on-one contact and we’re still on the Slack channel so that contact is continuing.”

Onehealthng.com using AI to speed patient support

Nigerian online pharmacist onehealthng.com uses AI to make sure patients can quickly contact the right consultant for medical advice. It hopes to extend and improve this service in the coming months.

Adeola Alli, founder and chief executive of onehealthng.com, appreciated the open and self-paced structure of the program. “I liked that you could just go in and select what you wanted,” says Alli. “I really liked the founders’ meetup. It was useful to hear from, and talk to, my peers. There was also great content on fund-raising strategies and product marketing.”

From the technical side, onehealthng.com lead engineer Samuel Tobi said sessions on optimizing processes and rapidly scaling systems were especially helpful. “We had one-on-one sessions with architects and AWS experts which were really useful,” he says. “The introduction to AI services helped us to work out our pipeline and how we can increase its use in the future.”

Adadot benefits from the EMEA Startup Loft Accelerator’s technical mentors

Adadot styles itself as a fitness tracker for development work. It measures collaboration, productivity, and wellbeing to help remote teams work better together.

Adadot appreciated working with AWS Startup Loft Accelerator because of its technical focus. “Lots of accelerators are all about fundraising but we didn’t need that,” says Adadot founder Alex Harris. “We’re a highly technical team, but it is very rare to get an incubator with that focus – our tech guys are often left on their own.”

Harris adds that the program brought a ”bit of sanity.” “It can be really isolating and lonely at a startup,” he says. “It was lovely to network with our peers but also to get more technical feedback on our product.”

In addition to technical and business advice and peer support, the program also provides the opportunity to join the AWS Activate program, which includes up to $25,000 in credits to spend on AWS services, a one-year free AWS Business Support subscription for up to $5,000, and access to the pre-built infrastructure templates.

Here’s to next year and more years to come!

We’d like to thank all the companies who have taken part this year. In 2023, we are hoping to make the courses even more tailored by streamlining companies by market. We will also be increasing the support activities we provide for companies which do not make it through the selection process.

Although the bulk of the program will remain virtual, we’re really excited to start in-person networking events in 2023, so startups joining the program could enjoy our unique community in real life.

Sessions for startups at re:Invent 2022

Alexa Korach — Thu, 27 Oct 2022 14:08:05 +0000

AWS re:Invent 2022 is fast approaching, and we can’t wait to see you in Las Vegas from November 28 through December 2! As usual, these four days will be jam-packed with activities, networking events, inspiring keynote speakers, breakout sessions that will allow you to connect with fellow attendees, interactive workshops, and more.

The best part? Every single event and activity is designed around–and geared toward—startup culture, making the entire conference 100% relevant to you.

re:Invent Attendee Guide

Serhat Can is the AWS Community Hero writing this year’s attendee guide. Serhat, who is based in Istanbul, is a co-founder at Resmo, a continuous cyber asset visibility and security solution for developers and world-class security teams.If he looks familiar, that may be because he speaks regularly at international conferences and organizes community events and meetups. He has also helped make devopsdays happen in over 80 cities around the world.

If you’re interested in seeing how AWS has revolutionized the game when it comes to building and scaling startups, we suggest you check out Can’s Attendee Guide.

Keep reading for an overview of this year’s highlights.

Content hub

With 50 content tracks and over 2,000 sessions offered in just four days, we know re:Invent can be a lot! That’s why we created content hubs, which offer livestreams of our most popular sessions.

Head over to any of our hubs—located at The Venetian, Caesars Forum, Wynn, MGM Conference Center, and Mandalay Bay—to chill and watch session streams while also grabbing a snack, sipping a drink, putting your feet up, or just taking a breather.

Breakout sessions

We all love an informative and rousing keynote address—but that doesn’t mean that the merits of a breakout session should be overlooked. With hour-long, lecture-style breakout sessions ranging from intermediate (200) through expert (400), we’re offering something for everyone. Stick around for the 15-minute Q&A at the end!

If you choose to attend a silent breakout session, you’ll be given headphones and welcomed into a room where each attendee can listen to whichever session they choose. There will be no Q&A at the end. Virtual attendees will have access to all of the following sessions on-demand a week after the conference ends.

Session ID	Title
STP201	Success secrets for mergers and acquisitions
STP202	Startups at scale: What to expect when building an engineering team
STP204	Refactoring the product and engineering relationship
STP205	A journey to increased developer productivity
STP206	Cloud strategies for equity valuation and investor readiness
STP207	Automating and evidencing key compliance security controls
STP307	Scaling a SaaS company for public company readiness
STP210	AWS optimization: Actionable steps for immediate results
STP211	Scaling on AWS for your first 10 million users
STP214	Making the right hiring decisions to build impactful products
STP306	Evolving your security capabilities through key growth stages
STP304	Habits to reduce your cloud costs
FWM309	Design to code: Build an app frontend UI in minutes with AWS Amplify Studio
FWM308	What’s new for frontend web and mobile developers with AWS Amplify
DEI207	AWS Impact Accelerator: Being the megaphone

The full re:Invent session catalog is now live, and reserved seating for in-person attendees is now available. Register here or log in to start building your schedule.

Pro tip: Use the filters to find tracks, topics (hint: startups), and services you can’t miss—or simply search for the session ID.

Chalk talks

If your idea of a good time is getting the opportunity to ask all your most burning questions, we’ve got you. Our chalk talks are highly interactive. Each one opens with a short lecture from an AWS expert followed by a longer Q&A addressing real-world architecture challenges. Every hour-long talk features a main speaker as well as two AWS experts who will come prepared to answer your questions and even illustrate problems via whiteboard as the talk goes on.

Session ID	Title
STP203	Security from day one: Establishing a baseline
STP212	Simple steps to prepare your environment for compliance
STP305	How to stay afloat in your data lakes
STP208	Designing resilient architectures for startups

Workshops

There’s nothing quite like collaborative, hands-on learning, and we’re excited to bring that experience to you with this year’s selection of workshops. Every two-hour-long workshop opens with a short lecture given by an expert before the participants are split into teams. Each group is then given a scenario and encouraged to work together, using AWS services, to solve the problem. Participants should bring their laptops and be prepared to engage with one another.

Session ID	Title
STP301	Startup security: Techniques to stay secure while building quickly
STP302	Startup optimization: Tuning app performance for max efficiency
STP303	Deploying a complete machine learning fraud detection solution
PRT234	How to build a complete application with deploy-on-commit methodology (sponsored by DevRev)

Bootcamps

Our Bootcamps, which are all led by AWS experts, will help you build necessary, practical skills fast. We’re offering an array of exam preparation bootcamps that will help you prepare for your AWS Certification exams, technical bootcamps that will help you learn in-demand cloud skills, and partner training workshops exclusively for AWS Partners.

Session ID	Title
TNC102	AWS Cloud Technical Essentials
TNC223	AWS Partner: Building data analytics solutions using Amazon Redshift
TNC222	AWS Partner: Building data lakes
TNC225	AWS Partner: Machine learning on AWS

Don’t miss re:Invent 2022

Our annual re:Invent conferences have been some of the most exciting, comprehensive, and cutting-edge events in tech for over a decade now—and we don’t plan to change that anytime soon. After a full four days of transformative talks, networking events, and workshops, you’ll leave Las Vegas inspired by the global cloud community, confident about building your future with AWS, and determined to rethink what is possible when it comes to cloud computing.

If you can’t join us in person, we’d love to have you attend Virtual re:Invent! You’ll get access to all of the keynotes and leadership sessions, as well as on-demand breakout sessions throughout the conference. Register now, either in-person or virtually, to attend our biggest and most all-encompassing cloud computing event.

Nandini Ramani – Vice President, AWS Monitoring & Observability, speaks at a leadership session at re:Invent

How Latino startup founders are advancing healthcare equity

Ricardo Moguel — Fri, 14 Oct 2022 12:38:51 +0000

As with all best forms of innovation, great ideas stem from true need. In Mexico, there is a need for an equitable, efficient, and sustainable healthcare system. Latino startup founders are addressing this need and advancing healthcare equity by leveraging artificial intelligence (AI) to drive better patient outcomes.

Zenda expands access to private medical coverage using AI

Zenda, a customized insurance startup for urban millennials and Gen Z, is working to reduce inequality in access to primary care due to lack of private health insurance.

Diego Muradás, Zenda co-founder and CEO

According to CEO Diego Muradás, more than 115 million people in Mexico don’t have access to private health insurance. “Our mission is to open access to affordable health services through our digital platform. We have been building not only a private, very accessible private health insurance, but also a health suite,” explains Diego. “This helps us to deliver basic health services like telemedicine or digital checkups. AI models, for example, help us to measure the biological age of people.”

Zenda uses biological age and other factors to provide insights into which types of illness they should focus on to maximize the effectiveness of their coverage program. Per Diego, “Our goal is not to cover everything—our goal is to cover the things that could happen to you with higher probability.”

Diego pointed out that a major challenge with training their AI models has been the inherent bias in most datasets. The team took great care in building a curated dataset tailored to the Latin American population. “Our dataset is specifically tailored for Mexicans. We are a very diverse society. Our model works well with the composition of Mexican society: the dataset incorporates ages, the genders, the skin color ranges. The model is something that we are very proud of.”

Diego says conversations and community building within the Amazon Web Services (AWS) startup ecosystem have been instrumental in product ideation and how to drive bigger impact. “AWS has given us the ability to build a community with not only them, but also within the broader community of entrepreneurship in here in Mexico,” he explains. “It’s especially helpful to be able to come to AWS solutions architects with a problem and say, ‘Hey, yeah, I would like to do this.’”

Clivi maximizes reach of doctors to treat diabetes in Mexico

Clivi is another Mexican startup that is addressing Latin America’s healthcare equity challenges, with a specific focus on the Type II diabetes.

Ricardo Moguel, Clivi founder and CEO

Founder and CEO Ricardo Moguel says, “Diabetes is the second leading cause of death in Mexico. Our mission is to bring people affordable care and help them to control their diabetes so these people can extend their life.”

Clivi collaborated closely with endocrinologists, nutritionists, and psychologists to build an AI-based application for improving patient outcomes. In the 12 months that the application has been live, the team has demonstrated a significant reduction in the A1C score of patients who have been in the program for at least three months. This drop continues throughout the product lifecycle.

Clivi’s key to success is maximizing the capacity and reach of individual doctors. “To treat diabetes in Mexico, there is one endocrinologist to almost 10,000 patients,” explains Ricardo. “What we are building is this tool that allows the doctor to generate more capacity.”

Clivi uses human-in-the-loop technology by using natural language processing techniques to predict a doctor’s responses to patient inquiries. The application can also automatically recommend adjustments to patient treatment programs for doctors to further augment their reach. “The application will select people that we need to evaluate, and then the doctor will decide whether these people need to adjust their treatments,” he explains.

With AWS solutions such as Amazon Comprehend Medical, the Clivi team expects to further boost the efficiency of medical providers to reach even more patients. “We believe that with AWS, we can improve the accuracy of our predictions and do it faster,” says Ricardo. “In the beginning, we were able to maximize the capacity of a doctor. Our next challenge is to go 10 times the original capacity in the next 24 months.”

Conclusion

With the support of AWS, startups like Zenda and Clivi plan to continue to create better health outcomes and a more equitable healthcare system. Ricardo believes the Latin American healthcare market is, “attractive to innovators who want to create a positive impact in people’s lives in Latin America.” As Diego says, “The way we use tools such as AI is super powerful nowadays and this gives us the possibility to do things that were not possible before.”

Related resources

How Pieces Technologies leverages AWS services to predict patient outcomes

Dr. Ruben Amarasingham — Thu, 13 Oct 2022 13:41:03 +0000

What would you do if you could predict the future?

With the advancing capabilities of predictive artificial intelligence (AI)/machine learning (ML) tools, forecasting the future—at least probabilistically—may be in reach.

Pieces Technologies, Inc. (Pieces), a healthcare and life sciences startup, is blazing a trail in the predictive AI/ML space. Pieces is a software as a service (SaaS)-based AI platform integrated into a hospital’s electronic health record (EHR). Their mission is to improve care by providing clinical insights along the patient journey. They offer predictions of health events such as projected discharge dates, anticipated clinical and non-clinical barriers to discharge, and risk of readmission, before they occur. Pieces also provides insights to healthcare providers in natural language, and optimizes the overall clarity of the patient’s clinical issues so care teams can work more efficiently.

Dr. Ruben Amarasingham, founder and CEO of Pieces Technologies

CEO and founder Dr. Ruben Amarasingham founded Pieces to help providers and patients achieve better outcomes. He explains, “Healthcare work is essential. Yet physicians, nurses, and other people are leaving medicine because of burnout and because of powerlessness. We see enormous technological sophistication in other parts of our lives—why can’t we apply that to medicine?”

During his 13 years as a hospital physician, Ruben saw that “there were a number of patients that kept being readmitted to the hospital. They would come into the hospital, receive treatment, and then one to three weeks later they would be back.” With post-graduate training in biomedical informatics and a focus on predicting clinical events, this trend inspired Ruben to help.

“I saw an intersection between my research background, my clinical practice, and what was happening in the hospitals.”

To apply his theory to a real-world use case, “We began to mathematically model—based on the characteristics of the patients—which patients might be readmitted. Then we applied clinical and non-clinical resources to the at-risk patient, including social determinants. And we found that we could identify and prevent re-hospitalization before it occurred.”

The ability to predict patient rehospitalization led Ruben to found Pieces in 2016. The scope of Pieces’ services for healthcare providers and patients has grown rapidly. Today, “We’re looking at any kind of patient adverse event and saying, ‘Are we able to predict it ahead of time?’ If so, we deliver an insight to individuals on the ground ahead of time in a way that they can act on it.”

To deploy and maintain their AI engine at scale, Ruben says Amazon Web Services (AWS) are integral: “We would not be able to do this work if we did not have an unbelievably dynamic, scalable cloud on which to do it. It’s not possible on premises. You need to have these modeling systems learn from large datasets and you need the ability to manipulate or modify the AI performance at scale and across clients.”

After a brief period during their incubation when Pieces used an on-premises environment, “We switched to AWS and we haven’t looked back. It’s been terrific. We are incredibly grateful for the AWS environment.”

The two most critical AWS services that Pieces uses to run their models in production and at scale are Amazon Managed Workflows for Apache Airflow (Amazon MWAA) and Amazon Elastic Kubernetes Service (Amazon EKS). Ruben explains, “We use Amazon EKS to run Kubernetes instances of Amazon Elastic Compute Cloud (Amazon EC2) and Amazon Relational Database Service (Amazon RDS).” The ability to monitor health system client sites in real time is critical for Pieces, so they use Amazon EC2 and Amazon OpenSearch Service to log analytics for real-time application monitoring.

Ruben is passionate about creating AI/ML predictions that are relevant and accessible. “We’re most excited about incorporating natural language generation into our software. That capability is the where Pieces is with AI right now.” Pieces leverages natural language generation to give healthcare providers predictions within context and in plain language. Natural language makes it easier for healthcare providers to evaluate the AI’s judgment for the most critical patient needs, while putting less critical predictions aside. This makes it easier to act on information in a timely and beneficial way.

“Natural language generation can help providers and patients by translating information—at scale—in the way that they can best absorb it. Whether that’s a clinical summary for a physician, or discharge instructions for a family member with, say, a seventh-grade education level.”

In the near future, we can expect Pieces to launch a hand-off summary for physicians that allows them to pass their patient to another physician if needed. The summary is generated using some of the clinical data that AWS solutions allow Pieces to collect. According to the Joint Commission, physician hand-offs are a significant problem in healthcare because “an estimated 80% of serious medical errors involve miscommunication between caregivers when patients are transferred or handed-off.” By automating this task, Pieces hopes to improve patient outcomes while also, “bringing joy of medicine practice into a system that is enormously complicated,” explains Ruben.

The Pieces Technologies team at their Dallas, Texas office

For startups that want to follow in Pieces footsteps, Ruben advises, “Healthcare is highly regulated, very precise, and you have to test things more rigidly than, say, a retail startup. It’s okay to take a slower course, and it’s important to find investors and backers that understand the healthcare environment and the benefits of incremental progress.”

The success is worth it, says Ruben, because, “If you can go through all of the steps, the impact is profound. You’re doing good for the world.”

Related resources:

Celebrating Hispanic Heritage Month with Hispanic Startup Founders on AWS: Part 3

Mauricio Di Bartolomeo — Fri, 07 Oct 2022 15:00:57 +0000

At AWS, we know that building a dream is best achieved by having support from people like you who want to get where you’re going as much as you do. To celebrate Hispanic Heritage Month, we’re recognizing the achievements and contributions of Hispanic and Latino Americans who have inspired others to achieve success.

This series of blog posts features Hispanic founders who share their unique challenges, aspirations, and how AWS helps them carry out their missions.

Ledn: Advancing equality through borderless banking

“Al mal tiempo, buena cara,” is a phrase familiar to many Latin American people during times of adversity: “To bad times, a good face.” Ledn co-founder and Chief Strategy Officer Mauricio Di Bartolomeo embodies this sentiment well. After living through political and financial instability in his home country of Venezuela, Mauricio set out on a mission to provide people everywhere—particularly in countries with instability—a better way to control their finances. His solution is Ledn, a financial services provider that leverages the reach of digital assets to serve clients globally, providing the same rates and level of service regardless of where they are located.

“I cannot disassociate my Venezuela story from my Bitcoin journey,” explains Mauricio.

Born and raised in Barquisimeto, Venezuela, Mauricio grew up as his country’s government sank into authoritarianism. At the same time, severe inflation led to financial instability. “Economics didn’t play out in Venezuela the way that the books said it should,” and Mauricio kept, “diving deeper into financial books, trying to gain a better understanding of the system’s flaws.” His interest in finance led him to attend Richard Ivey School of Business at Western University in Canada, where he earned an undergraduate and a master’s degree in Business Administration.

Mauricio’s involvement with Bitcoin arose from his youngest brother, who began Bitcoin mining in 2014 and wanted to remain in Venezuela. “My father sent me the Bitcoin whitepaper and said, ‘What do you think of this?’ I saw that it had a global market and the main input was electricity, which was heavily subsidized in Venezuela.”

Upon returning to Venezuela for Christmas, Mauricio’s “aha moment” happened while watching his brother convert Bitcoin into Bolivares. Inspired, he returned to Canada to build his own mining facility with his college best friend, Adam Reeds.

One problem Mauricio and Adam immediately saw in the community was that no Bitcoin business had access to financing.

“Our revenues were in Bitcoin, but our expenses were in dollars. No one wanted to sell their Bitcoin assets for cash. Banks and private lenders would say it was not an asset, and we could not borrow against it,” says Mauricio.

Adam Reeds, CEO and co-founder, and Mauricio Di Bartolomeo, CSO and co-founder, at the first Ledn office in Toronto

This is where the opportunity arose. “We saw a way to solve our own problem: By creating a service that would provide loans backed by Bitcoin.”

It was a problem they were equipped to solve: Adam had spent the last decade working with large pension funds to finance green energy infrastructure. He knew what investors wanted. Mauricio understood the Bitcoin landscape and the financial needs of Latin American people.

Together, they began to build Ledn.

“AWS played a big role in our early days, especially your AWS Activate startup program, because it gave us credits and the ability to test and use AWS services. It allowed us to build on a world-class platform, when at the time—frankly—we couldn’t have afforded to. AWS thought a lot about how to set up the program and what we would need, so it was very easy for us to test it. It was a symbiotic relationship—if it didn’t work out, then hey, it didn’t work out. And if it does work out, our company and AWS are both going to do very well together,” says Mauricio.

Founded in 2018, Ledn issued Canada’s first Bitcoin-backed loan that same year. In addition to loans, they also offer savings accounts that allow people to gain interest on their Bitcoin or US dollar-pegged stablecoin: USDC. Today, Ledn helps clients in over 130 countries access credit and savings products and bring their digital assets to life.

Mauricio is particularly proud of how they help Latin American customers. “We were the first lender to offer these services in Spanish. We bet very heavily in Latin America, early on. Everybody was trying to bank the unbanked with iPhone 8 apps and English content. We felt that was disconnected from their needs. Many unbanked people don’t have iPhones, they don’t speak English, and they are very nervous about new financial products.”

When asked if his Latin American identity influences his mission at Ledn, he explains, “So much. I think a lot of times, hard circumstances that we can’t understand give us lessons that we don’t know will be valuable later.”

“I love speaking Spanish. I love talking to people from back home. I love helping people back home. One of the biggest pains when I left was this feeling that I could no longer help. One of the things I love about Ledn is that I get to do that.”

To other founders looking to follow in his footsteps, Mauricio advises to have conviction in your mission and stay motivated: “There was a time in my life where I thought, ‘Maybe I’ll be a customer service rep for somebody.’ That’s all I saw when I went online and looked for banking opportunities for Spanish speakers. There’s never a C-suite position that says, ‘Speaking Spanish is required.’ But I wanted to put my understanding of Latin American problems to use. I wanted to help people.”

He continues, “When you are a founder, if your idea is innovative enough, most people won’t get it. To succeed, focus on the problem you are solving, understand exactly what your client needs, and deliver on the promises you made to them,” says Mauricio. “You have to be willing to invest your time, your money, and your reputation. If you follow through on these and have conviction, you’re going to be unstoppable.”

Adam Reeds and Mauricio Di Bartolomeo at Blockchain Summit Latam in Mexico City, Mexico

Ledn envisions their future impact as a vehicle for bringing people the ability to fulfill their goals. Mauricio hopes, “One day, people will say that because of Ledn, they are able to regain their financial independence. Governments should not be able to take away your right to convert your local currency into other assets and protect its value. So, I hope that because of companies like Ledn, people will say that they were able to regain control and fulfill their goals.”

“In five to ten years’ time, I hope that people in places like Latin America and Africa and all over the world are able to say, ‘Thank God for Bitcoin because it lets me choose where to take my money and where I feel safe.’”

We can expect a lot from Ledn in the next few years: More products that translate digital wealth into tangible assets—Bitcoin to buy a house, a car, groceries. They are also working on a suite of wealth management products for accredited investors. Mauricio sees unlimited potential in the synergy between, “North American investor markets with interest in generating income and Latin American borrowers. They pair and complement each other quite well. That is what we plan to help with.”

Special thanks to Alex Pajiltsev, Account Manager

Related resources:

Got a great idea? Build on AWS Activate

Angel Gambino — Wed, 05 Oct 2022 22:48:31 +0000

Calling all startup founders! Jump-start your idea using AWS Activate. Since 2013, AWS for Startups has offered industry-leading resources and AWS credits to over 200,000 startups, but that was just the beginning.

Starting today, from the moment you join Activate, you’ll also gain instant access to Activate Console, which offers a robust suite of benefits, including technical training, tested and proven infrastructure templates, personalized guidance, and more.

Previously, as a startup you had to apply for credits before you could access Activate benefits. Plus, as soon as you received Activate credits, the clock started ticking towards your credits’ expiration date.

Now, when you’re ready to deploy your solution, you can apply for Activate credits, unlock cost-optimization tips to help offset infrastructure costs, and access exclusive third-party offers from our trusted partners.

How to get started using AWS Activate Console

Step 1: Join AWS Activate

Using your AWS Account, complete the enrollment form. All startups can apply — even if all you have is an idea right now. If you don’t have an AWS Account, sign up today!

Step 2: Access your AWS Activate Console

As soon as you join Activate, you’ll have instant access to your personalized Activate Console. Think of the Activate Console as your Activate smart hub — it’s where you’ll access all the tools, education, and resources we’ve designed just for your startup.

More reasons to explore your Activate Console:

Build at your own pace. Prove out your minimum viable product (MVP) on more than 40 solution templates tailored to your use case. No infrastructure experience needed here — your Activate Console provides personalized recommendations and exclusive trainings to optimize your building experience.
Get personalized content that grows with you. Startups can be at different phases in their journeys. A founder who is tweaking a great idea needs different resources than startups who are celebrating their Series A funding. Your Activate Console delivers customized resources based on the stage and role you’re in, simplifying every step of your journey.
Learn with AWS experts. Get guidance from the experts who power the world’s best companies. Training sessions, interactive how-to’s, and specialized events help you build the foundational skills you need to level up your startup.
Real-time updates. Your Activate Console knows what your startup needs. It gives you real-time updates on the health of your AWS environment, helps you discover new AWS services, and is the monitoring tool you need to keep your growth on track.
Access to AWS Startup Lofts. Access to AWS Startup Lofts includes webinars, 1:1 sessions with AWS startup experts, and co-working space in select locations.

Step 3: Apply for AWS Activate credits

Depending on your startup stage, you may qualify for up to $100,000 in Activate credits, as well as additional AWS support credits. Plus, when you’re approved for credits, you’ll also gain access to:

Exclusive offers. If you have Activate credits already, you also unlock access to exclusive offers from trusted organizations such as Notion, Miro, Hubspot, and Stripe. The discounts, free trial periods, credits, and other startup-friendly perks are all designed to help you grow and succeed.
Cost-optimization tools. If you have Activate credits, your Activate Console will help you maximize your spend with a Cost and Credits Summary. It gives you real-time visibility into your balances, alerts you of pending credit expiration, and offers tailored cost-optimization tips to make the most of your spend.

Don’t rush this step — we want you to build at your own pace. When you are ready to deploy your solution, you can apply for the Activate credits that can help offset that cost.

Get started building today!

You’ve got the great idea. Let’s get started building it together, today. Join AWS Activate.

Scaling your early stage startup: Three award-winning startups share secrets of their success

Ina Stuve — Tue, 04 Oct 2022 16:40:25 +0000

There’s no blueprint for transforming an early stage startup into a global success story. Given how diverse tech startups are, it makes perfect sense that there’s no one-size-fits-all solution.

However, through analyzing the actions and attitudes of businesses that have made it to the big league, it is possible to identify a pattern of behaviors, actions, and attitudes that has fueled their growth. With this in mind, we asked three quickly growing startups from the AWS Software Startup Awards to share the secrets of their success. Here’s what they told us.

Customers can be your best allies

Having tripled its revenue for the most recent financial year and with a market value of more than $300 million, Dublin, Ireland startup Tines is thriving. It’s also just won Gold in the Rising Star category of the AWS Software Startup Awards.

Thomas Kinsella, co-founder and COO, Tines

Thomas Kinsella, co-founder and chief operating officer, attributes Tines’ success to its customers, who include Canva, Hubspot, Coinbase, and Auth0. He describes them as the company’s “best allies.”

“Getting support and feedback from our customers, whether that’s about adding new features, simplifying our platform or rolling out a new initiative, has really helped us push forward.” Thomas explains. “Listening and learning from our customers’ world-class security teams has been crucial for our success and will always be a priority for us.”

Aim high, but move fast

“Aim high, but move fast.” That’s the advice from Jem Walters, co-founder and chief technology officer of the UK startup money saving app Snoop. Since launching in April 2020, the app has been downloaded 700,000 times. The company scooped up Silver in the Rising Star element of the AWS Software Startup Awards, and was named Innovation of the Year at the 2021 British Banking Awards.

Walters points out that a great idea can’t be proven until it’s put in front of customers. “Get something simple up and running as quickly as possible—put it out there to start the feedback and iterate cycle,” he says. “Putting high-quality working software in front of your target users is the only way to really know if you have a viable business.”

Be ready to adapt

Turning a setback into an opportunity is the hallmark of a successful startup. Before the pandemic, PixelMax was using game engine technology and visual 3D software to create virtual versions of factories to monitor real-time data and improve productivity on manufacturing lines. The Manchester-based tech company had a global project lined up with an aerospace firm, but when lockdown kicked in and planes stopped flying, the order was put on hold.

To help it continue operating through remote working, PixelMax turned to AWS. The shift to the cloud proved to be the catalyst to completely pivot the business model to focus on a virtual, immersive 3D workplace and events platform aimed at solving the hybrid working challenge.

PixelMax, which took Bronze in the Rising Star category of the AWS Software Startup Awards, has seen significant growth since it reinvented itself. They now forecast growth to double each year for the next two years and its customers include high-profile global brands such as European football governing body UEFA.

AWS services turbo-charge growth

When it comes to streamlining processes and freeing up time and space for product innovation, Thomas says AWS Systems Manager has been the most impactful tool. With it, the Tines team manages its fleet of production machines quickly and easily and it saves them the hassle of setting up their own secure shell access.

Jem Walters, co-founder and CTO, Snoop

Snoop uses more than 50 AWS services, which Jem describes as “a classic case of the whole is greater than the sum of its parts.” The company has adopted a serverless approach, placing AWS Lambda, AWS Fargate, and Amazon DynamoDB at the heart of its platform.

“AWS services have been fundamental to our success story to date. Adopting a serverless approach means our engineers spend very little time on managing infrastructure and most of their time creating value for our business by building new product features, reducing risk, and optimizing costs.”

PixelMax, too, sees serverless tech as “extremely helpful” to its trajectory, particularly Amazon API Gateway and Lambda, which have enabled the team to rapidly build a highly scalable platform without needing to continuously manage and maintain infrastructure. PixelMax co-founder Shay O’Connell, says, “This allows engineers to spend more time focusing on more complex elements of our platform, such as 3D world streaming.”

A good support network is essential

Having people around to lean on and offer expert advice is invaluable. Joining the AWS Activate program “jump-started” Snoop’s journey, according to Jem, and the company is also benefiting from commercial arrangements with trusted partners through AWS Marketplace.

Tines joined the AWS Partner Network and also created an AWS Marketplace listing last year, unlocking a slew of go-to-market opportunities. Thomas reports that this has “streamlined procurement and created enhanced value for customers.” The Tines team is also making the most of co-marketing opportunities on offer through AWS, such as exhibiting at AWS Global Summits in New York, London, and Tel Aviv this year to boost brand awareness and make more connections across the cloud community.

On joining Activate, PixelMax received AWS service credits worth $110,000, which have allowed them to “accelerate our business, develop our software, and enable the business to produce a market-leading product” Jem says.

Focus on taking small steps to reach big goals

No one thing will make you successful, whether it’s choosing a new tool, launching a new marketing campaign, or hiring a great salesperson, but each of these things will make you one or two percent better, points out Thomas. “Continuously improving will enable you to build a successful business over time. Gauge your progress, celebrate every win, and acknowledge any failures so you know when it’s time for another round of goal-setting.”

Related resources

Startup Advocate: Build better business models with network effects

Angel Gambino — Mon, 03 Oct 2022 23:30:41 +0000

Network effects are fundamental to many business models, and virality can be crucial for certain aspects of growth of some businesses.

So, what are network effects?

At its most basic, network effects happen when the product or service that a company offers increases in value as more people use it.

Why are they beneficial?

Network effects—and there are many different types of them—are a primary way to create a competitive advantage. They’re also the best option when it comes to seeing long-term exponential growth.

What about virality?

Virality, on the other hand, occurs when your existing users get you new users for free. This usually happens when information about a company’s service or product spreads rapidly via word of mouth.

Why isn’t virality as helpful to growth as network effects?

Virality is exciting for obvious reasons. It feels good to see more people using – and talking about – your product/service. And it can be very effective and cost efficient for generating rapid awareness and adoption – important aspects of growth.

The downsides, however, are that for many business models it’s generally not great for increasing the value of the core customer proposition, it isn’t effective when it comes to overall defensibility, and the numbers often times aren’t great in terms of customer retention.

Types of network effects:

Broadly, there are four different types of network effects: marketplaces, interaction networks, data networks, and platforms. However, these can be broken down even further into 16 different types of network effects driven business models:

Physical (direct)
Protocol (direct)
Personal utility (direct)
Personal (direct)
Market network (direct)
Hub-and-Spoke
Marketplace (2-sided)
Platform (2-sided)
Asymptotic marketplaces (2-sided)
Expertise
Data
Tech performance
Language
Bandwagon (social)
Belief (social)
Tribal (social)

Success with network effects and the cold start problem

Many of the most famous tech companies and startups owe some of their success to network effects, including Twitter, Apple, Uber, Lyft, and Waze. For example, in the case of ridesharing like Uber and Lyft, the value of the marketplace improves and is interdependent as both sides of the market grow, whereas with Waze – the more quality real-time data points you get from a larger user base, the quicker and better the reliability of the user experience becomes, which builds trust in the brand. From these examples, you can see how scaling with network effects has improved the value of the user experience for these kinds of companies over time, but early stage startups who want to leverage network effects can find it challenging. We call this the “cold start” problem.

The cold start problem

The cold start problem is an issue that nearly every founder runs into. At the very beginning, how can you have network effects when you don’t yet have customers? Different companies have tackled it in different ways, but with the right plan, it is possible to start with zero customers and scale to billions using network effects. Andrew Chen of a16z talks about how different companies have done this successfully (and analyzes why others have failed) in his book The Cold Start Problem: How to start and scale network effects.

Virality and network effects at Gameplay

In 1999, I was on the founding team of Gameplay, Europe’s biggest online ecommerce store for games, gamers, and online multiplayer gaming. I was working for the founding CEO developing new online multiplayer gaming experiences and partnerships to accelerate growth. In order to gain attention and grow brand awareness amongst casual gamers, we created a “celebrity death match” between Robbie Williams and Liam Gallagher, who were publicly feuding at the time. It garnered so much press and web traffic that all of our servers immediately crashed.

While this was a great example of virality, it wasn’t necessarily the answer for Gameplay’s long-term growth. However, when it came to massive multiplayer games, many of the community features we developed at the time became intrinsic to the gaming experience, including tips and cheats and digital assets and league tables and enabling the developer community to build new patches and creatives to develop downloadable content/digital assets/skills in the marketplace we developed. This approach made gameplay as a company and gameplay as a gaming experience for hardcore gamers pivotal to our growth and expansion. For us, the viral games were a top-of-funnel way to lure a wide range of gamers to the site, and the network effects built into and around the games hooked them into spending hours with us, retuning repeatedly throughout the day, and becoming creators in the marketplace, which gave us a low-cost way to continually innovate and offer value to our users. Now, those were the heady days of the internet and I’m proud to say that although the business is different today, Gameplay survived through it all. We were too early in the market, but we did learn some early insights that highlighted the difference between the power of network effects and the useful but fleeting impact of virality.

My takeaway

At Gameplay, when we focused on what we now call network effects, as many game companies do now, we were winning. When we used virality to get people to the site and convert to our call to action (CTA) with a low customer acquisition content (CAC) and then layered different combinations and types of network effects, we yielded the greatest results. And if we had had access to Amazon Web Services (AWS)—and all of the infrastructure, resources, and expertise—at that stage, we would have been able to ensure that our entire network didn’t crash that day!

With virality, you have to constantly come up with new forms of viral engineering and content to stay ahead. This leads to geometric growth, which is great to see, especially at early stages and can impact the perceived value of your business—but it doesn’t necessarily lead to growth in value of your offer to customers. Determining how you can leverage network effects, so that the awareness and engagement you’re getting from viral growth can also contribute to increasing the value of your product or service, can quickly become a winning formula if you get it right.

Celebrating Hispanic Heritage Month with Hispanic Startup Founders on AWS: Part 2

Augie Del Rio — Wed, 28 Sep 2022 15:56:12 +0000

This series of blog posts features Hispanic founders who share their unique challenges, aspirations, and how AWS helps them carry out their missions.

Gallus Insights: Solutions built from entrepreneurial spirit

Founders are born from customers who dream of a better solution. Augie Del Rio, co-founder and CEO of Gallus Insights (Gallus), is fulfilling one such dream by providing middle market companies with advanced analytics services to support timely, informed decision making. Gallus provides tactical and strategic insights across financial, marketing, human resources, capital markets, and more for a fraction of the cost that it would take clients to provide their own analytics.

Per Augie, “We turn raw data into insights across the company: financials, marketing, human resources, operations, you name it. We help with business decisions by explaining what is happening to your business, why it’s happening, what could happen. Most importantly, what should you do about it?”

Augie’s path to CEO began in Monterrey, Mexico, where he lived until the age of 15 before moving to the US. He went to High School in Arkansas then attended the University of Amsterdam in the Netherlands, where he obtained a degree in Economics.

Augie Del Rio, co-founder and CEO, Gallus Insights

“Mexico is full of entrepreneurs. You can find artists, restaurateurs, greengrocers, and much more on every corner of Mexico City. And globally, there are major Hispanic companies like Sofktek and Kavak. Hispanics are an entrepreneurial bunch.”

After a retail banking career at HSBC, Augie returned to the US to earn an MBA from the Yale School of Management and worked at Goldman Sachs’ Investment Banking Division. After five years on Wall Street, he worked at Caliber Home Loans as head of finance. That’s where Augie had his “a-ha” moment.

“I realized middle market America had a big basic problem: turning data into insights. Data is created every second. There’s data on customer interactions, employee productivity, financials—just to name a few. This data typically sits in silos and is not readily accessible, combinable, shareable, and analyzable.” Augie says. “Consequently, many leaders end up making gut-based decisions… Or, others end up hiring armies of analysts to pull data manually. By the time they process the data and deliver it to decision makers, it’s stale.”

To quickly and efficiently put data in the hands of decision makers, Gallus technology brings data from various systems and databases into its centralized cloud data platform, which is powered by Amazon Web Services (AWS). The data is secure, and its design allows Gallus to leverage the full power of the cloud, which brings unlimited computing resources and unlimited amounts of data for top performance without any scalability restrictions. “The cloud is effectively your limit,” stated Augie.

With Gallus, mortgage companies have one source of truth and access to a plethora of analytical outputs that help executives make better decisions. They provide analytics at various functional levels (from the CEO office all the way to the loan officer) and at various subject levels (from pipeline analytics to recruiting status).

“I was talking to a client, and we were discussing the path his company could take. In a simple discussion he told me, ‘I need this amount of capital to get where I want to go,’” Augie explains. “I used Gallus, and could see that his capital requirement projection was very optimistic. To reach his goals, he would need ten times the capital that you think you need. In just one conversation, we delivered a quick insight into how to adjust his planning and create a realistic business plan.”

“We use AWS across the backend and frontend of our technology offering. We wouldn’t be able to service clients without it,” explains Augie. “We use Amazon Relational Database Service (Amazon RDS) to store data, AWS WorkSpaces Web to work in a secure environment, Amazon Simple Storage Service (Amazon S3) as object storage, AWS Amplify for our web development, Amazon Cognito for consumer identity management…and so on. AWS powers Gallus across our technology platform. And now, we are developing machine learning models through Amazon SageMaker. We are very excited about what we can aspire to accomplish with AWS.”

Augie mentioned he is very appreciative of the support he has received within the Hispanic community. “The level of support has been awe-inspiring. We have received significant help in the form of mentorship and networking support to even get clients. I’m proud of what’s going on in the Latin American startup world. We are launching phenomenal enterprises and, in doing so, disrupting industries.”

With plans to expand their product to other industries within Financial Services, Augie wants Gallus to become a one-stop-shop that optimizes the cost of data analytics while helping customers overcome the hurdles of fragmented data sources and systems. “Expanding into adjacent markets is the next step. We started with mortgage lending, and our next step is to use the power of our platform for banks and credit unions.”

For other founders who want to achieve similar success, Augie’s advice is two-fold. “Once you’ve identified a problem and you are passionate about solving it, have the conviction to go and do it. The entrepreneurial process itself—paired with passion and diligence—will give you the answers you need to successfully execute your plan.” He also advises, “There are a lot of benefits on the AWS side that I had no idea about: their entrepreneurship program has been super helpful and AWS Activate provides credits for startups. It’s like you help startup firms live rent-free for the first few months or years. Memo to startup founders: Knock on the AWS door. They will help.”

Special thanks to Ashley Wax, Demand Generation Representative

Related resources

AWS announces 25 startups selected for the AWS Impact Accelerator for Women Founders

Keely O'Neill — Tue, 27 Sep 2022 12:59:04 +0000

Fresh off the success of the inaugural Amazon Web Services (AWS) Impact Accelerator for Black founders, AWS is continuing its $30 million commitment to provide underrepresented founders with the resources, capital, and community they need to level the startup playing field.

Narrowing the gender funding gap

For our next cohort, AWS is working toward narrowing the gender funding gap of the startup world. Soon, 25 women founders will arrive at AWS headquarters in Seattle, WA, to kick off their intensive eight-week program. The finalists were selected from a competitive field of applicants and chosen by a diverse committee of AWS Startups experts based on the strength of their idea, technical readiness, and interviews with our team.

The 25 startups, by industry

We’re excited to help the following founders from across the US accelerate their businesses and to learn more about the wide range of ideas, backgrounds, and industries they represent.

Each of these startups will receive up to $225,000 in cash and AWS Activate credits, an extensive and individually curated training curriculum, mentoring and technical guidance, introductions to Amazon leaders and teams, networking opportunities with potential investors, and ongoing advisory support.

Additionally, the program provides each of these founders the opportunity to augment their network and foster friendships and partnerships with their fellow founders that will outlast the intensive eight weeks.

AWS is committed to delivering guidance and resources beyond the duration of the program and becoming a trusted partner as these women founders grow and accelerate their businesses. In the months and years ahead, the startups will have access to a virtual community, alumni events, curriculum (including future content), plus the opportunity to tap into ongoing support from program mentors and experts.

Please join us in extending a warm welcome to the AWS Impact Accelerator for Women Founders.

Industry: Education

Dreami | Founder: Ashima Sharma | Location: California

Dreami powers mentorship-driven workforce development programs for Educational, Non-Profit, Government organizations. The platform streamlines mentor-mentee matching, scheduling, training and tracking, to create high-impact opportunities for work-ready candidates.

Local Civics | Founder: Beverly Leon | Location: New York

Local Civics enables K-12 students and educators to build civic engagement and community leadership skills through a game-based digital learning platform. It teaches young people about power and how to harness their unique gifts to empower themselves and their communities.

TeeRead | Founding Team: Trinidad Bensusan Millé & Gabriele Battiato | Location: Florida

TeeRead is a platform for parents and teachers that uses AI speech recognition to make reading intervention scalable. The platform automatically diagnoses students’ reading levels, pinpoints reading issues and curates content based on a student’s ability.

Industry: Financial Services

Miren | Founder: Gabriela Campoverde | Location: New York

Miren provides community development financial institutions (CDFIs) and non-profit small business foundations with software that helps them engage with communities and mission-driven startups to distribute beneficial capital to some of the nation’s most underserved neighborhoods.

Totem | Founder: Amber Buker | Location: Oklahoma

Totem was founded and funded by Native Americans to create pathways to financial inclusion for Indigenous people through digital banking. Totem partners with sovereign tribal governments to provide relevant financial products and education, and to make tribal benefits more accessible.

Walkthrough | Founder: Megan Valentine | Location: Pennsylvania

Walkthrough gives users access to personalized financial advice to help them on their financial journey. The app incorporates behavioral science and gamification to keep members encouraged and motivated as they make decisions about investing, spending, wealth management and more.

Industry: Gaming

rrecess | Founder: Bee Law | Location: Oklahoma

rrecess is a timed social game with friends and “friends of friends.” The app connects people through shared interests and randoms, bringing them together at specific times of the day for meaningful interactions. rrecess represents a community-driven solution to social media fatigue.

Industry: Government

Cleare | Founding Team: Tisia Saffold & Ian Kiku | Location: Virginia

The childcare industry is at the forefront of our economy; however, 80% of daycares today operate without a viable tech solution. Cleare is compliance software for home and center daycares that streamlines the licensing inspection process and digitizes document collection.

Industry: Healthcare

EmpiricaLab | Founder: Norma Padron | Location: Texas

EmpiricaLab is a collaborative knowledge-sharing platform that enables peer-based training for healthcare teams. The platform connects team members with customized, relevant training content, empowering them to set, work toward, and achieve goals in a professional setting.

Euphoria | Founder: Kate Anthony | Location: Colorado

Euphoria helps transgender individuals navigate the arduous process of gender transition through a combination of four unique apps. Solace is a healthcare navigator app; Bliss helps financially plan for transition; Clarity is a sense-of-self tracker; and Devotion is a daily affirmations app.

Nessle | Founder: Carly Buxton | Location: Virginia

The Nessle platform helps new and expectant parents connect with perinatal experts nationwide. Through a highly-filterable, responsive web platform, parents can search, find and pay for virtual and in-person support from trained experts across the full spectrum of parenting concerns.

RxPlace | Founding Team: Patty Bollenbach & Franklin Williams | Location: California

RxPlace digitizes the process of medication procurement, while bringing transparency and efficiency to prescription fulfillment. RxPlace delivers a frictionless purchasing experience for pharmacists, so they can focus on clinical responsibilities and spend more time with patients.

Industry: Human Resources/Talent Acquisition

Culturora | Founder: Jody Madala | Location: Pennsylvania

Culturora is a networking platform that helps companies build a more inclusive workplace culture, to attract and retain top talent. The app engages employees and encourages them to participate in purposeful conversations using video technology and conversation prompts.

DIVERSD | Founder: Jennifer Williams & Jared Alessandroni | Location: New York

DIVERSD is an AI-driven approach to shaping Human Resources around the concerns of overlooked groups. A “CDO in a box,” DIVERSD offers a bot interface for nonjudgmental HR interactions, pattern detection to subvert marginalization and integrations to preserve anonymity.

Matchplicity | Founding Team: Lauren Burke Silva, Jovan Silva, Lori Burke, & Zita Steglich-Ross | Location: Washington D.C.

Matchplicity is decentralizing recruitment platforms by empowering niche, community-based organizations to mobilize their professional populations. The platform connects the most qualified candidates directly to hiring companies through an AI-powered matchmaking platform.

Industry: Media & Entertainment

ForUsApps | Founder: Saniya Shah & Omer Winrauke | Location: New Hampshire

ForUsApps gives content creators next-gen AI-powered tools to support startups and side-hustles. The company’s collection of technologies is accessible, practical and, most importantly, easy to use by creators and businesses as they seek to scale their social media presence.

Industry: Professional Services

Call Emmy | Founder: Arezou Zarafshan | Location: Colorado

Call Emmy is an on-demand marketplace that connects hyperlocal, qualified household and childcare service providers to families. The company seeks to solve some of the most persistent daily challenges working families face, while prioritizing accessibility, convenience and safety.

Clyn | Founder: Diana Muturia | Location: Arizona

Clyn is a SaaS-enabled marketplace platform for home upkeep and cleaning services. By combining a marketplace model and SaaS tools, Clyn helps service providers grow their scope of services to attract and retain customers, and gives them the autonomy to scale their business.

Dinner Daily | Founder: Laurin Mills | Location: Massachusetts

Dinner Daily offers personalized meal planning for consumers and strategic growth opportunities for grocers. The app solves the challenges of dinner planning by generating unique shopping lists specific to local grocers, while helping consumers save money and eat better week after week.

Industry: Retail

Glou | Founder: Karen Lee | Location: Massachusetts

Glou is the first consumer-to-consumer beauty marketplace, enabling users to buy and sell makeup, skincare and beauty products. The platform empowers beauty enthusiasts to become more conscious consumers, pushing against an industry that encourages over-consumption.

Industry: Software & Internet

Alta Tools | Founding Team: Son Ca Vu & Scott Ostler | Location: California

Alta is an AI-powered mobile platform that helps anyone design, build and launch mobile apps. With its simple no-code interface, powerful automation and smart design tools, Alta empowers businesses to quickly build and launch mobile apps across all major platforms simultaneously.

Monday Hero | Founding Team: Nazli Temurtas & Burcu Geneci | Location: California

Monday Hero is a low-code SaaS platform that helps startups, freelancers and mobile development companies create mobile apps 60% faster. The platform automatically converts Figma, Adobe XD, Sketch App and other design files to code, to expedite build times.

Standd | Founding Team: Julie Saltman & Stephen Solka | Location: Washington D.C.

Standd is a SaaS platform that gives lawyers visibility over their firm’s information, how it’s connected and how their team is using it. The platform integrates files with the data storage and research systems lawyers rely on, to enable quicker decision-making and better outcomes.

Industry: Sustainability

ECGO | Founder: Nicole Toole | Location: Georgia

ECGO rewards users for their recycling efforts. The app offers incentives to college and university students who choose to recycle, including discounts to popular restaurants and local businesses. ECGO aims to create a connection between recycling and reward at a personal level.

Industry: Travel

Inclusive Journeys | Founding Team: Crystal Egli & Parker McMullen Bushman | Location: Colorado

Inclusive Journeys helps people of marginalized identities find safe and welcoming spaces when traveling—and encourages inclusive businesses to open their doors to an even wider segment of customers. The app helps users locate, rate and review inclusive spaces wherever they travel.

Learn more about the AWS Impact Accelerator program

Celebrating Hispanic Heritage Month with Hispanic startup founders on AWS: Part 1

Israel Niezen — Wed, 21 Sep 2022 16:17:30 +0000

At AWS, we know that building a dream is best achieved by having support from people like you who want to get where you’re going as much as you do. To celebrate Hispanic Heritage Month, we’re recognizing the achievements and contributions of Hispanic and Latino Americans who have inspired others to achieve success.

This series of blog posts features Hispanic founders who share their unique challenges, aspirations, and how AWS helps them carry out their missions.

Factored: The transformative power of technology

CEO Israel Niezen co-founded Factored with a dream to help US companies bridge the data science talent gap while providing technology careers to top engineers in Latin America. That dream manifested into a business that accomplishes a dual mission. Factored develops advanced artificial intelligence (AI) and data science solutions for US companies. Per Israel, they also “employ engineers in Latin America whose individual economic impact leads to the creation of four-to-five additional jobs in local communities.”

Coming to the US from Peru with his mother and siblings at an early age, Israel quickly recognized the opportunity in the US for technically skilled workers to gain skills and earn livelihoods not available in many parts of the world. He relates to the hope inherent to the American Dream, while acknowledging that as an immigrant, “We’re underrepresented in technology and business leadership roles. I am acutely aware of how fortunate I am of having been able to benefit from legal immigration into this country and being educated here.”

Israel Niezen, CEO and co-founder

“Being a Hispanic American gives you the best of both worlds. It gives you the ability to match the technical and business rigor expected in Silicon Valley with the human warmth and passion that Latin Americans are known for. That’s a superpower that we have and should leverage more often. Make our Hispanic heritage front and center because, when well utilized, it’s a level of passion and energy that really, really motivates humans around a cause.”

Factored was founded in 2019, built from a wellspring of motivation and determination that started many years earlier. After studying finance for his undergraduate degree and earning an MBA at Harvard University, Israel spent two decades working at various technology companies where he learned programming and machine learning (ML). Whether by serendipity or sheer luck, his data science studies led to a correspondence with Dr. Andrew Ng, who shares Israel’s vision for investing in Latin American data science talent.

“We started with the hypotheses that Latin America had excellent engineers, but at the time not a lot of them had been exposed to Silicon Valley levels of rigor in terms of training or work opportunities,” said Israel. Dr. Ng had recently established the AI Fund, a prestigious venture studio that incubates new businesses that want to build the future with AI. They were coincidentally assembling a team around training the best ML engineers in Latin America. The fit was right, Israel explained, and “after many rounds of conversations, we decided to partner and build this company together.”

Pairing Factored’s work with Latin American talent and high-profile technology companies has also led to inspirational personal stories: “One of our engineers joined us from a very small company in Colombia.” Israel says, “We invested four months in full-time training before placing him in several projects for high caliber Silicon Valley companies. Thanks in large part to our training, exposure, and his experience at Factored, he got a fellowship at Stanford, and within three years of joining us, left to go to Open AI, one of the most prestigious AI research companies globally, proving that some of the best talent in the world just needs to be discovered and given an opportunity to shine.”

Israel attributes stories like these to the “equalizing power of technology.” Without it, “it is unlikely that either Stanford or Open AI or any other prospective company would have known that this engineer existed.”

Providing the talent and expertise of Latin American engineers to top Silicon Valley firms became a virtuous cycle. Today, three years after founding the company, Factored is an AWS Select Tier Services Partner and places their talent in AI and data science roles within industries such as financial services, healthcare, manufacturing, advertising, and retail.

“In helping transform our communities in the US and Latin America, we hope to also change the perception of Hispanics and Hispanic talent, specifically in technology.”

The way Factored leverages AWS has scaled as much as their business. Factored “became involved with AWS very early on, as part of the AWS Activate program, which was crucial to our development and exponential growth.” Israel says. “When you’re a startup and you don’t have a lot of funding, those credits and that training allow you to very quickly and cost effectively build proof of concepts, demos for clients, and gain the trust and ability to deliver value that led to opening further accounts with clients. I think engaging with Activate is something all startups should leverage early on.”

If you’re wondering what’s next for Factored, Israel says he hopes to become one of the best data science and data engineering companies in the world while creating employment and training opportunities for as many talented humans as possible. “We started by building ML teams and have now grown our data engineering and analytics services very quickly, while simultaneously certifying dozens of our engineers as AWS Cloud Practitioners and AWS Solutions Architects.” He mentions, “We hope to continue this fast growth, but continue without losing the magic of our culture that encourages our engineers to keep learning, dream of the possibilities, and find a way to make them real.”

He advises others looking to follow in his footsteps to prioritize execution, be extremely selective on who you bring in to share your dream, practice sustainable growth, and—most importantly—be authentic and transparent about your culture and objectives because it, “sets the tone that aligns employees, investors, and customers around your mission.”

The Factored team currently employs 150 engineers and counting

Special thanks to Aaron Melgar, Startup AI/ML BDM and Tarryn Heath, Partner Development Representative

Related resources

Learn to optimize your cloud spend with the new, free AWS Cloud Economics for Startups series

AWS Editorial Team — Wed, 07 Sep 2022 23:11:20 +0000

The AWS Training and Certification team is offering a new, free video series: AWS Cloud Economics for Startups.

Throughout the 15 lessons, you’ll learn from use cases of other real-world AWS startup customers that will help inform and guide your own particular journey.

What will I learn?

Objective: You’ll learn how to work with AWS from the earliest stages of your startup through maturity while optimizing cloud spend on a pay-as-you-go shared responsibility model.
Target audience: Anyone interested in learning more about the value proposition of AWS Cloud products and services and how to get started on their cloud journey.
How long will it take to complete: 1 hour

What’s on offer in this course

Successful Startups in the Cloud

In this lesson, we’ll discuss how the flexibility of the AWS Cloud can improve your business’s viability and how to boost speed, save costs, and improve agility.

Introduction to Cloud Value Framework

This lesson introduces the four key pillars of the AWS Cloud Value Framework:

Cost Savings – Run your infrastructure affordably using cloud-based technology services.
Staff Productivity – Boost efficiency with automation and the cloud.
- In the use case, we’ll show you how Reddit experienced hyperscale growth, and how AWS helped them migrate to a relational database service to reduce the burden on the staff.
Operational Resilience – Avoid and recover from outages.
- For our example, we’ll show you how cryptocurrency platform Coinbase reduced its risk profile. “Our core tenets are security, scalability, and availability,” explains Director Rob Witoff. “Just as a traditional bank would heavily guard its customers’ assets inside a physical bank vault, we take the same or greater precautions with our servers.”
Business Agility – Harness the AWS Cloud to stay competitive while adapting to customer needs and market changes.
- Because costs tend to increase along with the service footprint, MoneySmart, the largest financial portal in the Southeast Asia region, used AWS Cost Explorer, AWS Trusted Advisor, and other tools to optimize their Amazon Elastic Cloud Compute (EC2) instances (virtual computers on which to run applications).

Rightsize Your Services

This lesson shows you how to match EC2 instance types and sizes to your workload and eliminate any you don’t use, an ongoing process as your company’s needs change.

Make Your Business Elastic

In this lesson, we’ll discuss how to use Amazon EC2 to scale up and down, acquiring and releasing resources automatically—using (and paying for) just what you need.

For example, MoneySmart continued their cost savings by launching a non-working-hours policy that scaled down EC2 instances when the office was closed.

Select the Right Pricing Model

AWS pricing is like that of utilities—you only pay for what you use.

Select the Right Pricing Model shows how to choose between Amazon EC2 On-Demand Instances, savings plans, Spot Instances, and Reserved Instances.

In the use case, you’ll learn how MoneySmart crafted a mixed pool of Spot Instances for their Amazon Elastic Container Service (Amazon ECS) and Amazon Elastic Kubernetes Service (Amazon EKS) clusters, for example, while providing container-driven auto scaling and resource rightsizing.

Match Storage Type to Your Need

Analyzing access and usage patterns, Match Storage Type to Your Need describes ways to build multiple storage solutions into a well-architected system.

Design Infrastructure with Budget in Mind

New startups tend to focus on getting to market quickly. However, as a business starts to scale, controlling costs becomes crucial. The last thing anyone wants is to see hard-earned customer acquisition eaten up by unnecessary operating expenses.

Design Infrastructure with Budget in Mind provides tools such as the AWS Pricing Calculator to optimize cloud spend.

Ready to get started?

This course includes presentations, use cases, and helpful resources. Learn more about AWS Cloud Economics for Startups or sign up today!

From idea to app: build your MVP or prototype at the AWS Amplify September hackathon

AWS Editorial Team — Thu, 01 Sep 2022 15:41:19 +0000

Attention developers: Do you have an idea for a web or mobile app? Eager to take it to the next level? Bring your idea to life—join the AWS Amplify hackathon, hosted throughout September by Hashnode and Amazon Web Services (AWS), and start building your application.

The hackathon begins on September 1, 2022 (12:00 GMT) and ends on September 30, 2022 (23:59 GMT).

Winners will be selected by the Amplify and Hashnode teams, and the results will be announced during the third week of October.

What’s AWS Amplify?

Amplify is a set of purpose-built tools and features that lets frontend web and mobile developers quickly and easily build full-stack applications on AWS. With Amplify, you can configure a web or mobile app backend, connect your app in minutes, visually build a web frontend user interface (UI), and manage app content outside the AWS console. Ship faster and scale effortlessly—with no cloud expertise needed.

What do I need to participate?

Just an AWS account and your great idea!

Great, sign me up!

To find more details on how to enter the hackathon and start building your application, please read and follow the instructions here.

You will also find details on the hackathon process, evaluation criteria, and the many prizes you can win, including swag, AWS Activate credits, and up to $1,000 in prize money.

Need some help?

Join Amplify’s community on Discord to get assistance from thousands of developers and AWS specialists.

See you soon!

This is a terrific opportunity for you to learn new skills, take advantage of AWS tools and services to power your ideas, and show the community what you’re capable of.

We’re excited to see your ideas and projects. Good luck!

Other upcoming events

AWS Web3 Ready | September 15, 2022 | One-day virtual event showcasing the exciting possibilities of employing new technologies like blockchain, NFT and Metaverse, and DeFi.
Event series: Startup Academy | Ongoing | From tightening up your pitch deck to building your first MVP, early stage startups will learn the basics of launching and growing your business.
Event series: Startup Scale | Ongoing | For startups further along, this series dives deep on the challenges when scaling. With technical labs, founder talks, and investor trade secrets, access experts who have been through rapid scaling.

Letter to a future founder: Build community, boost business, and dream big with the AWS Impact Accelerator program

AWS Editorial Team — Wed, 24 Aug 2022 18:39:30 +0000

This post was authored by Aireka Harvell, founder of Nodat and inaugural member of AWS Impact Accelerator for Black Founders.

Dear future AWS Impact Accelerator founder,

Welcome to the AWS Impact Accelerator program! I’m Aireka Harvell, founder of Nodat. Not too long ago, I was standing right where you are—about to kick off my journey in the Impact Accelerator program. I know how overwhelming those first few days can feel. It’s my hope that in sharing some of my experiences, I can help prepare you for the next eight weeks and beyond.

I remember walking into the Amazon offices for my first week of the program. It was completely surreal to be a small part of the bustling atmosphere. Everywhere I looked, an Amazonian was opening a door or pushing an elevator button to an unknown location. And when the elevator door opened for my floor and I heard the DJ filling the room with so much energy, I knew the whirlwind application and interview process had led me to exactly where I was supposed to be.

From applicant to accepted

I’m getting ahead of myself! Let me back up and tell you a bit about how I was accepted into the Impact Accelerator program. I first heard of the program in a newsletter from digitalundivided, the leading nonprofit that catalyzes economic growth for Latina and Black women entrepreneurs. AWS’ program was the first accelerator option I had come across for Black founders in a while that offered more than mentorship. But just a few hours remained to apply. So, I clicked the link and worked on my submission late into the night to meet the criteria.

Among the 1,600+ applicants, I was chosen for an initial phone interview. A few days after that, I gave a 10-minute pitch with a 20-minute Q&A. Immediately after my pitch, I received an invitation to meet soon as possible. Not only would Denise Quashie, Head of Startup Programs and Accelerators, be on the call, but VP of Global Startups for Amazon, Howard Wright, would be joining the discussion. I called my advisors and got to work prepping for the final round of interview questions.

In the meeting, Denise and Howard wanted to know if I could meet them in Seattle in a few weeks. But it had not occurred to me that they selected us for the program. So, I answered yes confidently and waited anxiously for the more complex questions about our technology. Instead, Howard turned the camera to his entire team at AWS and said, “Are we going to invest in Aireka? If so, thumbs up!” Everyone put up their thumbs and said, “Welcome to Amazon, Aireka!” I gasped and cried!

Now that I’ve gone through the program and can reflect on those game-changing moments, I realize I got there because I wasn’t afraid to brag about myself. Make sure you do the same. No accomplishment is too small! At the same time, don’t be afraid to talk about your weaknesses and the things you don’t have together. When else will you have the opportunity to ask for what you need and get it?

First impressions and making connections

If you’re anything like me and my fellow cohort members, you may be a little too nervous to network at that first gathering. Here’s a pro tip: look for Derek Pham. Over the eight weeks of my program and beyond, he would become known as the “Ice Breaker King” and was our primary shepherd for this experience.

Derek Pham, the “Icebreaker King”

He helped us open up to our fellow founders and to the Amazon mentors and leaders we worked with, which is absolutely critical for success in the program. When you can be vulnerable right from the beginning and receive the uplift from your new community, a bond forms that can’t be broken. Your startup will be stronger for it.

Rubber, meet road

During our in-person meetings in the first week of the program, the CTOs for each startup pair with tech solutions analysts and executives from the AWS technical team. With this support, we accurately set up our architecture and secure our products on AWS cloud servers, switching to Amazon Elastic Compute Cloud (Amazon EC2) instances and Amazon Simple Storage Service (Amazon S3) servers. We also met with executives from LinkedIn who helped us improve our profiles and optimize them for customer acquisition and investor outreach. Since that session, my page has grown to 2,500 followers and over 500+ connections. The engagement on our company page has increased also. In addition, we met with marketing and growth experts who would later coach us on creating strategies for increasing customer acquisition.

After that first week, we hit the ground running for the virtual component of the accelerator. Don’t expect a once-weekly chat or a video you can have on in the background. Each week is intense in the best way, consisting of hour-long workshops at least three times per week, plus meetings with your business and technical mentors.

In the eight weeks of the program we increased our daily active users by 35%, downloads by 633% weekly, and boosted our revenue by 23%.

Be honest about your needs and goals in these meetings, and be open to the feedback you receive to accelerate your business. For instance, in my first meeting with Courtney Gras, AWS Startup Community Lead, she helped me understand how our key performance indicators (KPIs) and metrics were poor indicators of growth, and gave us the tools we needed to more meaningfully measure our growth and increase momentum. Thanks to her guidance, in the eight weeks of the program we increased our daily active users by 35%, downloads by 633% weekly, and boosted our revenue by 23%.

AWS goes above and beyond to connect you to experts who understand your industry, background, and goals. It’s up to you to take in their advice and act on it.

Graduation day

Before you know it, you’ll be at your graduation day, surrounded by the community you’ve spent the past eight weeks building. That community will be unmatched—so don’t be afraid to lean on it.

During my time, one of my fellow founders worked on the graphics for our website, and in turn I helped him with a few pointers for his pitch. I flew around the country to support and brainstorm with cohort members, and stayed active in our incredible Slack channel, which was constantly buzzing with interactions.

The last day of the program may be as overwhelming as the first. After eight weeks of an intense shared experience, saying goodbye felt a bit like letting go of your mom’s hand the first time you cross the street alone. It’s a bit scary, but trust that you’re now equipped with the tools you need to take those steps on your own—and that the weight of the AWS ecosystem is behind you, ready to back your every move. Their support, and the fellowship with your cohort, does not end on graduation day.

Making the most of your experience

As you embark on this journey, remember a few things. First, be authentically you. You might be the most out-there person in the room, but that energy may help someone else reveal more about themselves. Additionally, allow yourself to be fully immersed in the experience. The more you can contribute and receive, the stronger and more long-lasting a relationship you’ll be able to build with your cohort and your mentors. Set your intention for this time, and be open to where it might lead you.

Finally, remember that not every step in this journey is going to be a “yes.” But the experience of the AWS Impact Accelerator is here to remind you that you are worth it. You are not crazy for dreaming this big.

Best of luck on your journey,

Aireka Harvell

Related resources

AWS Announces 25 Startups Selected for Inaugural AWS Impact Accelerator for Black Founders
Inside the Inaugural AWS Impact Accelerator Program for Black Founders
Applications Are Now Open for the Impact Accelerator for Women Founders
AWS Activate offers startups free tools, resources, and more to quickly get started on AWS

High-Growth Innovation Powered by Technology

AWS Editorial Team — Thu, 11 Aug 2022 19:30:17 +0000

Meet the startups that cut the edges of what technology can do and fly into innovation space in their solution-powered rocketships.

Most startups want to achieve the title of ‘high-growth’. It’s a moniker that defines their stratospheric growth within their sector. That labels them as a disruptor and change maker. And that paints them as innovative and interesting—as the kind of company that investors pay attention to and customers want to adopt. Fast-paced, exciting and sitting right on the very edges of what technology can do, these startups ignore the expected in favor of what can be, what might be. They are the next Etsy, Netflix, Airbnb, Uber and Calm, just to name a few high-growth startups that took the globe by storm.

Which is exactly what Signal AI, Synthesia and TrueLayer—all winners in the Rocketship category of the AWS Software Startups Awards—have done. These high-growth startups have applied the potential of technology to the needs of their industry and created solutions and platforms that are transforming how consumers and enterprises engage, transact and analyze. Signal AI won Gold, Synthesia won Silver and TrueLayer walked away with Bronze in an award ceremony that recognized their use of innovative technology solutions and AWS tools to achieve high-growth standards.

Gold medallist Signal AI started life in a garage and now has offices in London, New York, and Hong Kong

Describing itself as a ‘decision augmentation company’, Signal AI takes data and applies machine learning (ML) to create actionable insights for 40% of Fortune 500 companies. Today, Deloitte, Bank of America and Google use the platform for real-time market and media intelligence to uncover trends, risks and opportunities while supporting critical decision-making.

“We now have more than 700 customers around the world and employ more than 200 people,” says Luca Grulla, CTO of Signal AI. “We’ve raised just over $100 million in funding so far, and have been named one of the UK’s fastest growing technology companies in the Sunday Times Tech Track.”

“We drove the business forward by being ahead of demand, using a platform mindset and moving at pace.” – Luca Grulla, CTO of Signal AI.

Signal AI runs thousands of ML models over millions of documents every day at impressive speeds—the P90 of document enrichment latency is less than 30 seconds. Translated into layman’s terms, that’s the ability to sequence enough documents at a fast enough pace to maximize data output without compromising on quality or information value. The company has gained impressive ground since its inception, but the road to success hasn’t been without challenges. As Grulla points out, “Being disciplined in identifying the most impactful market or use cases to go after is always challenging for scale-ups. When building our business to business (B2B) platform, we had to always be ahead of the demand curve and drive the business forward using a platform mindset and moving at pace.”

The company has been flexing AWS capabilities to build a platform that ingests and enriches documents at scale using vertical scaling with diverse Amazon Elastic Compute Cloud (EC2) instance families and Amazon Aurora and vertical scaling patterns such as autoscaling and shading. The different storage options provided by AWS allow for Signal AI to leverage the storage it needs on demand with a mix of Amazon Aurora, Amazon ElastiCache and Amazon DynamoDB to handle low-latency lookup and data lakes built on top of Amazon Simple Storage Service (S3).

“AWS has been critical to our success, allowing us to scale quickly and experiment with different models and datasets,” says Grulla. “The multiple data analytics tools available in AWS allow us to maximize the value of our data and to support both our product and corporate data needs.”

It’s easy to see why this high-growth startup won Gold in the AWS Awards, especially as the team and leadership are passionate about the work that they do and the ways in which they can use technology to achieve their goals. When asked what piece of advice he would give to any other entrepreneur setting out on the high-growth road, Grulla said: “Find ways to experiment and learn what the market needs and then be laser-focused on establishing your position and market differentiation.”

In just five years, Silver winner Synthesia has created the world’s largest AI video platform

“Today, consumers expect video and audio content, and they take that preference with them to work, so we created a solution where anyone can generate video directly from their browser to meet this rapid demand for scalable content creation solutions,” says Victor Riparbelli, CEO and co-founder of Synthesia.

Synthesia operates the world’s largest AI video platform and generated more than 7 million unique videos for thousands of clients in the last year alone. The company’s client base ranges from global corporations to small businesses—all using Synthesia to produce uniform, on-brand video content for everything from internal training to personalized sales prospecting. Some of the clients currently on the Synthesia roster are EY, who uses the platform for client communications, and WPP who uses the platform to create personalized training videos for 50,000 employees. The company also powers celebrity campaigns such as PepsiCo’s Messi Messages and Malaria No More with David Beckham.

“Traditional methods of producing video and audio are expensive, cumbersome and require specialists,” says Riparbelli. “It’s impossible to keep the content library up to date as the business changes and the effort required is enormous. Synthesia closes that gap. Now, anyone can generate video without the need for cameras, studios or actors.”

Users simply select an avatar on the Synthesia platform, type in their script and generate a video within a few minutes. It’s a transformational approach that allows anyone to produce uniform, on-brand content and create a highly personalized AI avatar.

“Ultimately, we’re an AI company, and we’re building for a future where anyone can create Hollywood-grade video on a laptop,” says Riparbelli. “And on our way there, we’ll be solving some of the hardest and most fundamental problems in AI and computer vision.”

Synthesia has been using AWS since February 2018, only a few months after its inception, and continues to use the technology to drive its existing product development. In 2021, it significantly increased its AWS usage with GPU instances for Synthesia Studio—its ML platform.

Bronze medallist TrueLayer routes half of all open banking traffic in the UK, Ireland, and Spain

The platform processes billions of pounds in payments, and saw a 400% growth in payment volume and 800% growth in monthly payment value in 2021 alone.

“We enable anyone in the world to pay, onboard, and share financial data from any bank account simply, safely and within seconds,” says Francesco Simoneschi, CEO and Co-Founder of TrueLayer. “This means that, instead of paying with a card, customers can pay with their bank app in just a few clicks. No more keying in card details or trusting a website to store credentials.”

Today, millions of consumers use TrueLayer to pay for goods and services, whether investing through Freetrade, banking with Revolut, or buying a car on Cazoo. The company now supports some of Europe’s most innovative brands—high-growth powering high-growth—and has taken on more than 100 clients in the past 12 months.

“We are backed by leading investors that include Tiger Global, Addition, Stripe, Anthemis Group, Connect Ventures, Mouro Capital, Northzone, Tencent and Temasek,” concludes Simoneschi. “To date, we have raised $270 million and are valued at more than $1 billion.”

“We enable anyone in the world to pay, onboard and share financial data from any bank account simply, safely and within seconds.” – Francesco Simoneschi, CEO and Co-Founder of TrueLayer.

This true unicorn is smart at business but equally smart at technology, leveraging several AWS solutions to achieve its digital goals. Currently, the company runs its service on multiple Kubernetes clusters hosted on AWS along with Spot Instances on the AWS Elastic Compute Cloud, and DynamoDB.Spot Instances have enabled the company to be flexible in what it runs and for how long so it can dramatically reshape infrastructure to suit customer needs. DynamoDB is scalable, easy to administrate and cost-effective and benefits the company in running high-performance applications without worrying about underlying infrastructure or maintenance.

The AWS Software Startups Awards recognize innovative startups and entrepreneurs across several key categories, including Sustainability, Go-To-Market Innovator, Founder of the Year and Rising Star. Read about the other winners here.

Inside the Inaugural AWS Impact Accelerator Program for Black Founders

AWS Editorial Team — Wed, 10 Aug 2022 13:01:04 +0000

Brace for impact! The AWS Impact Accelerator program got off to a flying start—and it’s only the beginning of what this initiative has to offer.

It was a summer of Impact for AWS Startups. For eight intensive weeks, an incredible group of 25 Black-led startups became the first cohort of the AWS Impact Accelerator program, designed to provide underrepresented founders with the training, resources, and investment needed to accelerate their business.

Based everywhere from Compton to Baltimore and representing industries ranging from crypto to healthcare, our first cohort of Black founders are now better equipped to succeed in a startup space that has historically shut them out.

Now, AWS is ready to provide those same resources to our second cohort, which will focus on women founders. Think your woman-led startup could benefit from the program, or know of one that could? Here are some of the ways the AWS Impact Accelerator can help you accelerate in just 8 weeks:

Personalized coaching with a technical and business advisor: The AWS Impact Accelerator essentially adds a CTO to the startup’s team—one with in-depth knowledge of AWS and specialized experience based on the startup’s industry and goals. In week one, the advisors, made up of former founders, investors, and VCs, sit down with founders to identify key goals to accomplish in the next eight weeks. Startups meet with this mentor weekly, and get the hands-on technical and strategic help they need to accelerate their business.
$225,000 in capital: AWS is giving each startup $125,000 in zero-equity cash to scale their business, plus an additional $100,000 in AWS Activate credits to use as they build on AWS.
Amplified media reach: Our PR team helped our Black founders get their messages into local and relevant markets. Hundreds of news stories were produced about our first cohort, directly leading to investor and VC outreach.
A unique opportunity to augment their network through a lasting community: Impact Accelerator startups are chosen for their individual strengths. But when drawing on the expertise and talent of their cohort peers, each startup can get even more out of the program. Through dedicated Slack channels and virtual social events like wine tastings and trivia nights, these founders are building communities that will outlast the eight weeks of the program. In the first week of the Black Founder cohort alone, thousands of Slack messages were exchanged. Participants created specialized channels and went out of their way to help their peers—one non-technical founder immediately found a solution from a tech expert, and another organized in-person meetups. Whether it’s through cross-industry partnerships, a network they can always tap for answers, or lifelong friendships, participants are creating a community that will enhance their startup in the months and years ahead.
Unparalleled access to the Amazon ecosystem: During the first cohort, we made countless personal introductions between Black founders and top executives in Amazon business divisions like retail, music, and healthcare, as well as with trusted external contacts with investors.

That access is a critical benefit that will last far beyond the intensive eight-week program, and is one that other accelerators simply can’t offer. Investors who never would have answered cold calls or emails from these startups have started paying attention after AWS has made the intro, says Denise Quashie, Head of Worldwide Startup Marketing Programs at AWS.

“It’s an enormous privilege to be able to put Amazon’s connections and capabilities to work on behalf of these amazing startups,” Quashie says. “If we can open just one door for each one of these startups it really puts them on much equal footing with a those who are considered the majority and have those doors open much longer.”

And we’re not stopping here. AWS is eager to extend that level of access to our second cohort of women founders. If you’re interested in being part of the Impact Accelerator program or know a great candidate, find out more here.

Applications Are Now Open for the Impact Accelerator for Women Founders

AWS Editorial Team — Mon, 08 Aug 2022 14:01:09 +0000

AWS is launching a one-of-its-kind program that gives more women founders the support they need to accelerate their businesses.

Female founders receive just a portion of the funding that male founders do (some of the highest estimates say that just 18% of VC funding goes to startups with women and men on their founding teams, with only 2.8% and dropping to teams of only women founders). Even when women-led startups are funded, their battle is uphill compared to male founders, as they face challenges including a lack of available mentors and sponsors, biased sampling from early adopters, and the constant toll of microaggressions in a white and male-dominated space.

That’s not just bad news for women who want to build a startup—it’s bad news for everybody. Women are the driving force behind a thriving economy. AWS recognizes the immense need for an increase in inclusivity in the startup ecosystem. That’s why we are excited to be launching applications for the second cohort of our Impact Accelerator, exclusively focused on women-led businesses.

What Is the Impact Accelerator?

The AWS Impact Accelerator is a $30 million fund that provides Black, Latino, women, and LGBTQIA+ founders with equitable access to training, mentorship, tools, and resources. Over the course of eight weeks, 25 startups accelerate growth by developing their ventures alongside AWS technology experts, investors, and partners. From customer discovery and retention to architectural deep dives and Amazon learnings such as Working Backwards, founders participate in workshops, roundtable discussions, 1:1 mentoring, peer learning, and inspirational leadership chats from executive leaders. To further their growth, startups receive a $125,000 cash grant and $100,000 in AWS service credits—all at zero cost and for zero equity.

Focus on Women Founders

Throughout the program, women founders will learn from well-established experts from across the startup business and technical spectrum. Speakers and special guests for the Impact Accelerator for Women Founders will speak from their own lived experiences as overlooked and underestimated founders. Participants in the Women Founders program will learn from partners like Advancing Women in Tech, Visible Hands, and AWS partners like DocSend, Hubspot, Carta, and Brex. They’ll also benefit from sessions and modules curated specifically to address the challenges of women-led startups—each led by experts that have been through similar struggles and can help them accomplish their goals.

“The amount of tangible, measurable support that this Impact Accelerator has provided to my startup has far exceeded my expectations. My technical advisor helped me and my tech lead solve a problem that we had been toiling away on for weeks. My business advisor introduced me to marketing experts, people in my target market who I could conduct user interviews with, and others who have provided immense value. The entire accelerator has been a phenomenal experience.” – Alyse Nicole Dunn, CareCopilot, Cohort Member of Impact Accelerator for Black Founders

How to Apply – Five Key Requirements

Are you interested in applying to the AWS Impact Accelerator for Women Founders? Don’t wait! Applications are open August 8 – August 26, 2022, with limited space available for only 25 women-led startups. Submit your application here, and don’t forget to review the FAQ and terms and conditions for the full list of qualifying criteria.

Want tips to make your application stand out? Join this webinar from the Impact Accelerator Program Team to hear more about the application process, understand timelines and expectations for the interview process, and gain key insights. Plus, submit your questions in the chat—the team will be answering all of your remaining questions, live.

Top Five Criteria Reminders

Women-led – You must have one (1) or more CEO co-founder(s) who, alone or in the aggregate, own a majority (>51%) of the Company and identify as a woman.
CTO in place – You should have your CTO or technical lead in place, and be sure to include all founders in your video submission.
Based in the U.S. – Your founder(s) and company must be U.S. based—international startups will not be considered at this time.
Early funding stage – Your company hasn’t raised more than $500,000 in institutional funds and is no more than five years old.
Launched or fully developed product – To get the most from the program, founders need to be beyond the ideation stage—that means having a fully developed minimum viable product (MVP) or launched product.

Apply Today

The AWS Impact Accelerator gives high-potential, pre-seed startups the tools and knowledge to reach key milestones such as raising funds or acceptance into a seed-stage accelerator program, while creating powerful solutions in the cloud.

Applications are now open for Women Founders seeking to accelerate their startup’s growth from 0 to 60 in eight weeks. Apply online today.

Behind the Scenes at AWS Summit Toronto and Collision Conference 2022

AWS Editorial Team — Tue, 02 Aug 2022 18:00:11 +0000

What do you get when you mix an AWS Summit with one of the fastest growing tech conferences in Toronto? We call that one of our busiest weeks yet. Packed with networking, collaborating with peers, and learning from top thought leaders in the space, our week at AWS Summit Toronto and Collision Conference 2022 last month was a blast!

At the AWS Summit, Miriam McLemore, Director of Enterprise Strategy and Evangelism at AWS, shared her expertise on the innovations of the AWS platform and how AWS solutions are helping companies all over the globe. And over at Collision Conference 2022, Hollywood actress Lupita Nyongo graced the opening day stage with an appearance to discuss Super Sema, the first African American superhero animated series and what attracted her to the role. We also spotted “Dude With A Sign” with, well, a sign, and Batman flying around the Brampton Investments booth.

AWS is running events and Summits all over the globe that help technical professionals stay on the leading edge. And if we still haven’t seen you in person this season, here are just a few reasons to attend an AWS Summit:

Connect with and learn directly from AWS experts. Attend lectures, chalk talks, and panel discussions featuring industry-leading builders and founders who have been where you are.
Network with other startup founders. Recharge your phone and you social network at the AWS Startup loft, and get one-to-one advice from Startup Solutions Advocates.
See the latest AWS innovations and solutions. Get hands-on demos, meet with product leaders, and understand your startup options.
Be inspired by industry experts and leaders.

Visit aws.amazon.com/events/summits to learn more and register for an AWS Summit event near you!

ICYMI: Get Web3 Ready with AWS, Prevent Unexpected Costs, Celebrate Diversity, and More

AWS Editorial Team — Mon, 01 Aug 2022 18:47:38 +0000

Each month, the AWS Startups blog is packed with announcements, resources, stories, and videos. Let’s get caught up on anything that you might have missed. Then, head to Twitter and LinkedIn to share your reactions, and tell us what you’d like to read.

Announcements

Register Today for AWS Web3 Ready on September 15
Calling all founders, developers, and executives ready to explore what’s next in Blockchain, NFTs, cryptocurrency, and beyond. AWS Web3 Ready is a one-day virtual event, designed with the Web3-curious in mind. Get a first look at the speaker line up, and register to receive our exclusive NFT artwork.

Startup Spotlight

Founder Spotlight: Xata Co-founder Monica Sarbu Celebrates the Power of Diversity
Serial entrepreneur Monica Sarbu explains why celebrating diversity and empowering people is the key to success. Learn how her commitment to building with diversity from Day One has guided startup Xata to a $5 million funding round.

How Timehop Developed the World Class Ad Platform Nimbus with Support from AWS
We sat down with the Timehop and Nimbus CEO, Matt Raoul, David Leviev, VP of Programmatic Product Development, and Mark Laczynski, Senior Cloud Architect to discuss how the obstacles the company faced using third-party ad serving platforms led to the in-house creation of Nimbus and how they leveraged AWS solutions to not only improved the quality of ad servicing to their users, but also increase overall ad revenue.

Helpful Resources

Demystifying Startup Jargon
Is your head buzzing with buzzwords? Let’s unpack the terminology and the acronyms that define startup speak, so that any founder can talk entrepreneur and investment with ease.

What Funding is Best for My Startup: Bootstrap or Venture Capital?
When it comes to funding a business, two of the most common options are to bootstrap or to find venture capital (VC). The two different approaches can often be used in conjunction with one another to take a startup from one stage to the next. Learn the difference, and hear from a founder making that choice for his own startup.

How to Prevent Unexpected Costs for Startups While Building with AWS
As a startup, you want to focus your time on innovation and product market fit—not worrying about underlying infrastructure or unexpected costs. By following these four recommendations, you can extend your runway and create a long-term strategy for cost management.

The 10 Cs Super-Successful Startup Founders Have in Common
The qualities that make an uber-successful tech startup founder are a complex, even mystical, blend of traits. As we celebrate the stars of our AWS Software Startup awards, meet three incredible founders who share similar key characteristics and are blazing a distinctive trail.

Demystifying Startup Jargon

AWS Editorial Team — Fri, 29 Jul 2022 17:02:59 +0000

“There’s a lot of language that can really intimidate people and put them off. There needs to be a simplification of language in the funding sector to make it more accessible, especially around SaaS metrics and finance.” – Zandra Moore, CEO at Panintelligence.

Entering into the entrepreneur space can be daunting, especially for entrepreneurs who don’t have a grounding in funding, finance, and investment. They need to run the gauntlet of investment and finance, while also climbing the very steep learning hill of acronyms and industry-specific terminology. From non-dilutive capital to uncapped convertibles, shareholder agreements and dry powder, there’s a lot to learn, and often in a very short period of time.

Here you’ll find some of the most common startup terminology to help you navigate conversations across funding and finance with vocabulary that will keep you in the loop, and ahead of the game.

Accelerator: Centers or hubs where entrepreneurs can gain access to mentors, funding and support when they’re in the early stages of their business. Often, accelerator hubs are targeted at specific niches or industries, and can provide companies with an essential boost when they’re struggling to scale.

Angel Investor: Private investors with personal wealth that focus on financing small business ventures in exchange for equity. They make money on the percentage of the company that they own and their earnings or investments can be combined with venture capital funds or syndicates, which will all share a percentage of the profit.

Angel Syndicate: A group of individual angel investors that combine resources and strategy to invest into early-stage startups. Their goal is to pick a winner and share in the profit. The lead of the syndicate determines the carry (see below).

Bootstrap: ‘Pulled up by the bootstraps’ is a saying that’s been around since the 1800s to describe a person who has taken themselves to the next level and elevated their role, standing and financial status without any outside help. It’s got the same meaning for startups and finance—the entrepreneur has used their own funding to get their business off the ground.

Burn Rate: This term is used to determine the pace at which a company is using its capital to finance its overheads before it starts to see a positive return on the investment, or positive cashflow. It’s really the speed at which a company is losing money and is determined by subtracting expenses from revenue.

CAC: One of many acronyms that defines startup language, CAC stands for ‘customer acquisition cost’ and is used to determine your sales and marketing deliverables and value adds. You calculate it by adding up the costs of your sales and marketing, then dividing this number by the number of new customers you’ve gained over a specific period, usually a month.

Carry: Short for carried interest—the percentage of profit paid to the lead of a fund, angel syndicate or special purpose vehicle (SPV). The most common percentage used by the lead when launching a fund is around 20%.

Cap Table: Cap table is short for capitalization table and refers to the equity capitalization for a company. It’s a tool that’s often used for early stage companies and startups to list all the securities that the company has issued, such as stocks or ownership percentages.

Convertible: Convertibles allow for investors and entrepreneurs to value the company at a later date in the future. This is a clever way of helping early startups to gain access to funding while still uncertain of their valuation, or if they’re at a stage where it’s just too soon to get an accurate valuation. Convertibles can be capped or uncapped. A capped convertible places a limit on the valuation point at which investor notes convert to equity; with an uncapped convertible, the investor has no guarantee around the equity that their funding purchases.

Dilutive and Non-Dilutive Capital: When an investor offers dilutive funding, they are asking for a slice of your business that can include anything from control of decision making within the company to a share of future profits. Non-dilutive funding is the opposite—the investor is not expecting you to hand over parts of the company in exchange for their investment; they just want to see some return on their investment and for your company to shine.

Dry Powder: Sassy startup slang that defines how much cash an investor has available in reserve. It’s often used by VC investors or private equity companies that want to get the most, well, bang for their buck.

Low Hanging Fruit: A term often used across multiple industries, low hanging fruit refers to the simplest and easiest route to investment success for your business. It means you pick the partners and financing solutions that are achievable and that help you meet your business objectives as fast as possible.

MVP: MVP is short for minimum viable product and describes the earliest stage that your product or service can be sold to customers. Originally defined in the book Lean Startup by Eric Reis, the term has caught on and is used by many startups to get, well, started.

Private Equity (PE): A PE company is focused on investment into the private equity of startups using a variety of different strategies. These can include VC, growth capital and leveraged buyout. The company profits from the fees they charge their investors in the fund, and they hold onto this profit by making sure that they consistently create profit for investors by increasing the value of the companies they buy so they can sell them at a profit.

Special Purpose Vehicle (SPV): An SPV is a company formed as a subsidiary with the sole goal of taking on a particular business strategy, goal or activity. These are often seen in finance as a way of separating the parent company from the risk, while allowing for the company to still potentially benefit from its investment.

Term Sheet: Includes all the information an investor needs to know, and really understand, about making an investment with a specific company or investor. It will outline the details of the investment and should cover essential details such as the initial purchase price, assets, company valuations, investment amounts, stake percentages, voting rights, liquidation requirements and more. An investor will want to fully understand the term sheet before they sign.

Venture Capital (VC): A type of private equity and financing offered to startups that have proven, or expected, long-term potential. The funds come from investment banks, financial institutions or investors.

A final thought

These are some of the most common terms used in the startup world, but they only really scratch the surface. For any entrepreneur setting out on the journey across the finance ocean, it’s worth spending time unpacking the terms that best fit strategy and investment planning so that terminology doesn’t build a barrier between your startup and sustainable success.

Become a member of the AWS Startup Loft to listen to our podcast with Zandra Moore, CEO of Panintelligence, to discover the world of startups.

What Funding is Best for My Startup: Bootstrap or Venture Capital?

AWS Editorial Team — Tue, 26 Jul 2022 16:25:45 +0000

When it comes to funding a business, there are several options you can choose from, but two of the most common are to bootstrap or to find venture capital (VC). The two are very different approaches and can often be used in conjunction with one another to take a startup from one stage to the next. However, the type of funding route you choose will entirely depend on the type of business you have, the amount of money you have available, and your access to VC investment opportunities.

First, the difference between the two.

When you bootstrap your startup, you’re following in the footsteps of a saying that originated sometime in the 1900s. It comes from the saying ‘to pull yourself up by your bootstraps’ which means ‘to improve your quality of life without any support’. It’s defined by single-minded startup development that’s not reliant on anyone else for funding or growth—only you, the entrepreneur.

When you look for VC support for your startup, you’re looking to use private equity and financing that’s offered by investors to companies that they believe have potential. The type of investor and the shape of this investment can vary. Investors can be banks, well-off individuals, financial institutions, limited partnerships or any combination thereof, and they provide funding in exchange for equity. However, not all forms of VC are financial—you can also get technical support or managerial expertise as part of this investment role.

Bootstrapping and venture capital are very different approaches and each ask entirely different things of the entrepreneur. The list below unpacks some of the most important considerations that you need to make in order to choose between a bootstrapping or VC or a combination of the two.

“I’ll be very honest, this is my fourth startup and I wasn’t planning to do another, it’s hard.”
– Osvaldo Spadano, Founder and CEO of Akoova

1. Independent decision making

When it comes to deciding which direction your business should go in, what markets to enter, what products to develop, which people to hire, and what salaries to pay, bootstrapping gives you complete independence. As your own primary investor, you get to make every decision on your own. On the flip side, VC investment is more exacting as your investors will expect a measure of control and influence within your business as they’ve put their money into your vision.

Choosing between the two can come down to this very important point because some people want the control while others want the growth.

2. A single-minded focus

When it comes to VC investment, you will need to be prepared for a very different approach to business. With a VC on board, you will have to meet with your investors, focus on growth, and maintain solid relationships with the right people. This may sound like a drawback, but the value of VC lies in how it adds credibility to your business and the connections that your investor can help you make. You just need to make sure that you build a strong relationship with your VC to avoid unnecessary misunderstandings around the future of the business and to manage the balance between investor demands and your own vision.

Bootstrapping, of course, means that you can focus on whichever part of the business you most want to expand, grow or change. You can put profit at the heart of your business; you can manage your spend how you want; and you can put your customers at the heart of every interaction. These last two factors can sometimes get lost in a VC-funded startup as entrepreneurs are torn between so many different responsibilities and directives.

3. Support and connections,

When you bootstrap, you’re at your own financial mercy. You have little funding to reinvest into your business at first and you’re going to have to compromise on profit levels. Another downside of bootstrapping is that it can limit your competitive advantage as a result of this limited funding—you can’t try out new markets or experiment with new products or scale your workforce at speed. You may also find it more difficult to make the right connections and get specialized support compared with those who have opted for the VC route.

VC funding may come with terms and conditions, but it does also come with street cred, expert help, connections and instant networks. It’s in your VC’s best interests to connect you to the right people and make sure you’re getting all the right attention, so they’ll put time and effort into helping you build networks of your own.

4. The personal risk

There is no guarantee that any startup will succeed; many fail regardless of how they were funded. While there are many factors that influence a startup’s success story, there is less risk to personal funding and savings if the business goes through a VC. When you bootstrap, you may be using personal funds or savings to get your company off the ground so if it fails, you will lose all that money. VC gives you more financial freedom to fail, but the type of investment and agreement will influence how the failure will hit your bottom line.

That said, both methods of building a business are proven with many companies standing tall on their boots and many equally rising to prominence through VC. The choice often comes down to the personal expectations of the entrepreneur and the type of business they want to build. There’s no wrong answer and there is no reason why a company can’t turn to VC after a bootstrap start to kickstart its next phase of growth. The choice, as they say, is yours…

Whether you’re bootstrapping your startup with the Activate Founders program or have VC investment and are leveraging Activate Portfolio, we have a program that can support you.

You can also discover more about bootstrapping a successful startup by listening to Osvaldo Spadano, Founder and CEO of Akoova, on our podcast. Become a member of the AWS Startup Loft to learn more.

How to Prevent Unexpected Costs for Startups While Building with AWS

AWS Editorial Team — Fri, 22 Jul 2022 14:59:49 +0000

By Xuan Gao, Startup Solutions Architect, and Carole Suarez, Startup Solutions Architect

As a startup, your days are filled with competing priorities: you want to focus your time on innovation and product market fit—not worrying about underlying infrastructure or unexpected costs. The cloud management firm Right Scale estimates that wasted cloud spend averages about 35%, with small and medium-sized companies overspending the most. With this in mind, we’ve created this how-to guide to make sure your startup doesn’t end up spending thousands of dollars due to a spike that could have been prevented through monitoring or an alarm. We’ll cover best practices for new and existing AWS accounts when it comes to fundamental security, monitoring, and cost management. We’ll also get into the nitty gritty when it comes to proactively setting up alerts on anomalous usage of services due to over provisioning of services and/or misconfiguration. By following these four recommendations, you can extend your runway and create a long-term strategy for cost management.

Securing your account and workloads

The principle of least privileges is a security principle that entails granting only the permissions through Identity and Access Management (IAM) required to complete a task. Prohibiting unauthorized use of services allows you to more closely control access to your AWS Account and may prevent unauthorized charges. To do this, you’ll need to decide which users or applications will perform a specific task and the exact permissions needed to complete it. For instance, under the principle of least privileges, Business Intelligence analysts would only be granted the access to analytics services, such as Amazon QuickSight and Amazon Athena.

Regularly rotating access keys is another way to prevent extraneous charges. Make sure that all of your account users—including yourself—regularly change their passwords. This will limit both the amount of time a compromised credential can affect your startup and ensure that users will not be able to access resources after they’ve left the company.

Set up a multi-account AWS environment with AWS Organizations. This free AWS service will allow you to set up separate development and production accounts. As your workload grows and becomes more complex, you can remain flexible when it comes to billing, security controls, and budget requirements. We recommend dividing your workloads into production, testing, and development environments to more easily determine your operational costs, based on the regulatory and budget needs of your startup. This may also protect your production environment from unauthorized testing, which could lead to downtime or configuration errors.

Monitoring cost and usage

Use AWS Budgets to set up daily budgets and alerts. Daily—as opposed to monthly or weekly—granularity will alert you to upticks in charges via email and SNS. Daily alerts are especially useful if services are provisioned over the weekend. Daily alerts also allow you to manage resources appropriately, correct any misconfigurations before too much time has passed, and prevent unpleasant surprises at the end of the month.

Use AWS Cost Anomaly Detection to detect anomalous usage of services automatically. AWS Cost Anomaly Detection is effective in part because it’s customizable. By segmenting spends—for instance, tracking AWS Lambda and Amazon Simple Storage Service (Amazon S3) separately—you get fewer false alerts. You can also choose the level of granularity used to analyze spending. Our founders have already found this useful; using Cost Anomaly detection, one customer noticed a 50% uptick in Amazon CloudWatch spend and was able to quickly get in touch with their AWS team and better understand their usage of services.

Use AWS CloudTrail to set alerts for service usage. AWS CloudTrail is an easy-to-use tool that allows customers to review account activity and categorize it into “events”—essentially keeping a record of all related activity, which can help when troubleshooting the root cause of anomalous costs. For example, a user may have unauthorized access to a service and using AWS CloudTrail, you would be able to determine who, when and at what time the event occurred.

Use Amazon CloudWatch to set monitoring alerts for high usage. Through Amazon CloudWatch, you can enable billing alerts, create billing alarms, and receive SNS notifications when spending exceeds your threshold. Another benefit to Amazon CloudWatch is monitoring usage patterns over time and setting alerts, giving much-needed data when it comes to forecasting trends and optimizing spending going forward.

Leveraging AWS for additional support

Set up AWS Trusted Advisor for further security and cost optimization. Trusted Advisor is a resource that provides you with a full set of checks that become available as the support plan is upgraded. For example, when Amazon Elastic Block Stores (Amazon EBS) are created, customers incur a charge—even when they’re unattached or have low write activity. By implementing checks on Amazon EBS volumes, customers can determine which are under-utilized and remove them to cut costs. Similarly, Savings Plans will check customers’ usage of Amazon EC2, AWS Fargate, and AWS Lambda over a 30-day period and provide recommendations for usage amounts for one- to three-year periods—at a discounted rate.

Going forward with confidence

For startups looking to prevent unexpected costs while building on AWS, we highly recommend following the above steps. By securing your account, using monitoring tools to more closely manage service usage, and leveraging AWS Trusted Advisor, you’re on your way to better financial management. Our Well-Architected Framework Review can help you implement a clear-eyed strategy for success, providing insight on how to architect for best practices and maximize cost optimizations. For customers who have already discovered cost spikes, we recommend reaching out to the AWS team for support. You can view more information on support plans here.

Register Today for AWS Web3 Ready on September 15

AWS Editorial Team — Thu, 21 Jul 2022 14:20:11 +0000

Calling all business decision-makers, operators, founders, developers, and C-level executives who are ready to explore what’s next in Blockchain, NFTs, cryptocurrency, and beyond. AWS Web3 Ready is a one-day virtual event taking place September 15th, 2022, designed with the Web3-savvy and the Web3-curious in mind.

AWS is committed to supporting those who have adopted Web3, as well as those who are interested in learning more about this new phase of the internet. At AWS Web3 Ready Day, you’ll learn how Web3 companies are already leveraging a number of AWS products and services, and how AWS supports customers with infrastructure, speed-to-market, and visibility leveraging AWS’s core competencies of Compute, Storage, and Networking.

This jam-packed day will feature breakout discussions, panels, and fireside chats exploring how community and identity will be at the center of Web3—and how its mass adoption can benefit your startup. Our expert speakers and panelists will dive deep on the big questions: How can we use Web3 to democratize opportunity and ownership of the internet? How are community-led efforts—and the diversity of these communities—catalyzing Web3? How can we use Web3 to build more robust startups? And what are the benefits of using Web3 for your startup?

Our all-star lineup of panelists and speakers include Amy Wu, Managing Partner at FTX Ventures, Baron Davis, NBA All-Star and Founder of More Than Us Venture, Brit Morin, Founder of BFF, Mo Shaikh, Co-Founder of Aptos Labs, Janine Yorio, Chief Executive Officer at Everyrealm, and many more. With a wide range of experiences and expertise, these leaders will share their stories from the trenches—including decisions they made that helped them get where they are, what they wish they’d done differently, and how they anticipate the future of business changing as Web3 is mass adopted. We know you’ll want to be in the audience as they share their experiences and answer your questions.

Register for Web3 Ready Day and when you register, you can opt-in to receive our exclusive NFT artwork created by Shaylin Wallace. We can’t wait to see you there!

The 10 Cs Super-Successful Startup Founders Have in Common

AWS Editorial Team — Mon, 18 Jul 2022 14:00:27 +0000

The qualities that make an uber-successful tech startup founder are a complex, even mystical, blend of traits. It’s true there are many other factors which determine whether or not someone will hit the big time, but no matter what the sector, product, or service, we’ve found all super-star founders share key characteristics and attitudes. Here are our top 10 Cs that super-successful tech startup founders all have in common.

1. Communication

The ability to communicate well is vital, not just with investors but with customers, suppliers and the wider community and network. It’s also crucial to build a team of people who individually bring something fresh to the table while understanding each other’s viewpoints.

Maya Pindeus, Co-Founder and CEO of Humanising Autonomy and winner of AWS Software Startup Founder of the Year, gives her key ingredient for success as “a great, visionary and, most importantly, diverse team”.

Growing from three founders to more than 50 employees has been one of the toughest but most exciting challenges of the journey so far, according to Maya. She adds: “We’re lucky to have brought together a group of talented, motivated, and fun people.”

2. Collaboration

Trailblazers instinctively build a network of like-minded people to bounce ideas off and turn to for advice. It can be lonely running a business and sometimes impossible to see the bigger picture, so forging relationships with others at a similar stage in their journey can be invaluable. AWS Software Startup Founder of the Year Silver medalist Tanya Field is Co-Founder and Chief Product Officer of Novatiq Solutions. She advises finding the right partners to make your vision a reality as “now, more than ever, great businesses are based on winning partnerships.”

3. Conviction

When progress is slow, or it feels like you’re going backwards, successful founders keep pushing on. They understand that ideas and products need to keep evolving and if you’re not adapting and moving forward, you’re making life easier for your competitors.

Osvaldo Spadano scooped the Bronze medal in the AWS Software Startup Founder of the Year Awards. He says his toughest challenge since co-founding his business Akoova, which provides Magento cloud hosting on AWS for retailers, was “staying positive during the first year of bootstrapping the business, without a salary and with a wife and two children to look after.”

4. Competitive

Top founders know when to pivot and many, such as Akoova’s Osvaldo Spadano, favor a ‘lean thinking’ mentality. This means starting small, staying flexible and evolving constantly. Rapidly adapting to customer preferences is something Humanising Autonomy’s Maya Pindeus also believes in. She explains: “I believe in experimentation and an iterative design process when building a business. It’s so important to test your ideas early. “Make sure you talk to potential customers, users, and investors early, and listen to their feedback, iterate, and repeat, creating a fast-moving, exciting, and rewarding process.”

5. Competence

Efficiency and attention to detail might sound dull, but they’re absolutely crucial, especially when it comes to product quality and customer service. There are no shortcuts, which is why the most successful founders definitely do sweat the small stuff. Tanya Field points out: “It’s important to remember that the best idea in the world is useless if it is not executed correctly.”

Tanya’s tips include raising more money than you think you’re going to need and never underestimating how long it will take to achieve your goals because “excellence takes time.”

6. Can-do attitude

There will always be setbacks and obstacles that are outside a founder’s control, but what sets the elite few apart is their resilience. Not only do they bounce back, but they learn from their mistakes.

7. Call for help

Great founders don’t try to build or do everything themselves—they’ve learned to delegate or ask for help. It’s easy to get sucked into working 24/7, but there’s plenty of research that shows too much stress or overwork leads to bad decisions. Be big enough to recognize when someone else might be able to do it better, find the right angle, or come up with a smarter idea.

8. Cloud power

If you’re on a big mission and dealing with a lot of moving data, you need a trustworthy and secure place to store information, but it also has to be simple to use and easy to access. Maya Pindeus, who’s worked with AWS since launching Humanising Autonomy in 2017, says: “Having AWS cloud storage services has really allowed us to focus on providing our behavior AI products to customers in various fields—enabling the fast and easy deployment of our technology—rather than having to worry about the reliability of our storage platform.”

Novatiq Solutions’ Tanya Field adds that great customer service is crucial: “Having a dedicated manager to help us get the most from our AWS investments has been hugely valuable.”

9. Courage

Being bold enough to take a risk and try something new or unusual can be a game-changer. This means holding your nerve and trusting your instincts. Disrupting a sector by doing something in a completely transformative way often ruffles feathers and can feel like a lonely place to be.

10. Clear narrative

Create a compelling story for your brand from the start. It’ll help you to stand out and differentiate you and your business from competitors, especially when it comes to investors, customers, and the media. That said, don’t ever be tempted to bluff—keep it authentic or you might regret it at a later stage. Tell your story so that it reflects everything that’s unique about you and what you’re so determined that your startup will achieve.

The AWS Software Startups community gives you 1:1 access to software startup experts, exclusive startup events, and resources to help you at every stage of your journey. Join the AWS Software Startups community.

Founder Spotlight: Xata Co-founder Monica Sarbu Celebrates the Power of Diversity

AWS Editorial Team — Fri, 15 Jul 2022 13:54:07 +0000

Berlin startup Xata has blazed a trail since it burst on the scene in December 2020, thanks to its ultra-fresh take on managed databases. CEO and co-founder Monica Sarbu believes this success is driven by the talented and diverse team she’s brought together from a variety of backgrounds, cultures and countries.

Each colleague’s unique experiences and perspectives when combined in a single project create a dynamic and exciting workplace culture where people are inspired and excited to bring ideas to the table. It’s this creative vibe that’s helped Monica and her team build a serverless product that’s so user-friendly that it’s transforming the way developers and non-tech people set up and manage databases.

Xata is not Monica’s first entrepreneurial venture. She co-founded open-source project Packetbeat in 2013, which was snapped up by Elastic—the iconic software crew behind open source stack Elasticsearch, Kibana, Beats, and Logstash.

During her five years with Elastic, Monica was based in Berlin while managing four teams working from the US, Canada and Europe. She was also part of the core team who took Elastic to IPO.

How to hire the best talent in the world

One of the biggest lessons Monica took from Elastic was the upside of supporting team members to work from wherever they feel most comfortable and productive. This is reflected in the way she runs her own company, with team members based across Europe, and soon the US and Canada.

She explains: “It means you can hire the best engineers in the world—you’re not tied to a city or even a country. And when you have the best people in the world working on the same project, you build a really successful product.”

Driven by diversity to make a difference

It was a search for diversity that sparked Romanian-born Monica’s idea to start Xata. Having heard accounts of discrimination and burn-out, she set up a non-profit organization Tupu.io to mentor women, people of color, and other under-represented groups in the tech industry.

Monica searched for a serverless database to store details of all the mentors and mentees, but when she wasn’t able to find an easy-to-use option that ticked all the boxes, she realized she’d stumbled across a gap in the market.

Using representation to bring richness and depth to a business

A team that is diverse brings far more innovative solutions to the table, and this translates into amazing products, as Monica points out.

“If you have a team of software engineers who’ve all been to similar schools and worked at similar companies, they’ll have similar experience, outlooks and solutions,” she says.

“But when your team is truly diverse, they each bring a different perspective and fresh ideas to the table.”

And much like AWS, Monica is committed to celebrating differences from Day One. She points out that it’s important to think about building diversity into a team from day one. “Don’t wait until it’s a 50-person organization and then start trying to be diverse, as the longer you wait, the harder it will get,” she cautions.

Backers with a broader vision

Xata raised $5 million in September 2021 and plans another fundraising round later this year. Monica has worked hard to make sure Xata has angels and investors from a wide variety of backgrounds. Although she acknowledges that doing so takes more time and effort to achieve, it brings its own rewards, including wider networking opportunities and a more diverse bank of mentors.

All-embracing technology

As an entrepreneur who builds from scratch and moves fast, Monica believes technology support is crucial. Most of Xata’s customers run applications in the cloud, so she needed a cloud services provider that can guarantee low latency and network traffic costs. She immediately chose AWS.

“It was an easy decision,” Monica recalls. One of the things she rates highest about AWS is the sheer breadth of products and services available. She explains: “As an early-stage startup, you want to outsource as much as possible. You don’t want to manage services yourself, especially if you’re a small team, as you need to be free to concentrate all your efforts on building your product. This is very important, and it’s where AWS provides something really valuable.”

Activating the benefits of AWS

Monica is also a big fan of AWS Activate, which gives startups a wide range of benefits, including credits worth up to $100,000, one-on-one support from AWS experts, and guidance to help grow the business.

Monica has also been impressed by the level of customer service she’s encountered from the AWS team. “They not only help with product and services but also with go-to-market strategies and advice on how to become a successful company,” she says.

Amazon Elastic Container Service (Amazon ECS) has been particularly useful, as it’s helped the team to keep things simple. “Probably at some point, we’ll need to switch to Kubernetes but we’re trying to postpone that moment as long as possible because Amazon ECS is so easy to operate,” she adds.

Building a future that’s based on belonging

Meanwhile, Monica plans to keep building more products with the aim of expanding into a public company much like Elastic or MongoDB. She’s also equally determined to continue with her passion for mentoring others, via Tupu and giving back to the tech community. Inspiring others comes naturally to Monica.

With that in mind, one of her main pieces of advice to all would-be entrepreneurs is: “Have the courage to start! If you fail, you’ll have learned a lot and be a better person. “You’ll regret it far more if you never try to realize your dreams.”

Listen to our podcast with Monica Sarbu, CEO and co-founder of Xata, to discover the world of startups.

ICYMI: Meet the Inaugural Cohort of the Impact Accelerator, Come with Us to AWS Summit Atlanta, and More

AWS Editorial Team — Fri, 01 Jul 2022 13:46:10 +0000

Each month, the AWS Startups blog is packed with announcements, resources, stories, and videos. Let’s get caught up on anything you might have missed. Then, head to Twitter and LinkedIn to share your reactions, and tell us what you’d like to read next month.

Announcements

AWS Announces 25 Startups Selected for Inaugural AWS Impact Accelerator for Black Founders

In April, we announced a 3-year, $30 million commitment to the AWS Impact Accelerator, a series of programs designed to help high-potential, pre-seed startups led by underrepresented founders succeed. The first of these programs, the AWS Impact Accelerator for Black Founders, invited applications from U.S.-based startups to an overwhelming response, with applications pouring in from across the country. Learn more about the 25 startups selected.

AWS Expand Helps International Healthcare and Life Sciences Startups Develop Go-To-Market Strategies for the US Market

Amazon Web Services has partnered with MATTER, the premier healthcare technology incubator and innovation hub, to launch AWS Expand. The program supports startups entering the US market from Europe, the Middle East, and Africa (EMEA) by providing expert guidance, mentorship, client connections, an extensive network within the healthcare industry, and more. Learn more about how AWS is supporting healthcare startups.

Startup Spotlight

The Sustainability Equation: Why Tech for Good Means Tech for the Future

We need startups with innovative mindsets to step forward with solutions that will enable the world to achieve the Sustainable Development Goals (SDGs), startups that put sustainability at the heart of their technology, which is precisely what Basecamp Research, BioSimulytics Limited, and Sellalong Ltd. have done. Learn more about the 2022 AWS Software Startup Awards medalists in the Sustainability Champion category.

Helpful Resources

Video Series: The Startup Guide to AWS Services

The Startup Guide to AWS Services series consists of seven short videos and blog posts featuring AWS Solutions Architects, with related links and training videos. Each presentation highlights the advantages to working in a cloud environment, with brief examples that pair services to use cases.

Two Awesome Days at AWS Summit Atlanta

We’ve recapped our time at AWS Summit Atlanta, and if you weren’t able to attend, we have a behind-the-scenes look at what you missed. We heard from experts in AI/ML, Analytics, Business Intelligence, IoT, Databases, and more who connected and networked with industry peers to help them take advantage of the latest AWS solutions. Watch the video and see an AWS Summit through our eyes, and then register to attend the next event near you.

AWS Expand Helps International Healthcare and Life Sciences Startups Develop Go-To-Market Strategies for the US Market

AWS Editorial Team — Thu, 23 Jun 2022 16:10:09 +0000

The partnership with health tech incubator MATTER is the beginning of a larger campaign to help startups expand to new geographies.

Amazon Web Services has partnered with MATTER, the premier healthcare technology incubator and innovation hub, to launch AWS Expand. The program supports startups entering the US market from Europe, the Middle East, and Africa (EMEA) by providing expert guidance, mentorship, client connections, an extensive network within the healthcare industry, and more. For the 38 startups selected to participate in this first cohort this summer, Expand promises incredible opportunities for growth and development.

“This initiative is focused on our EMEA healthcare and life sciences startups that want to expand to the United States,” says Jared Saul, Global Head of Healthcare & Life Sciences Startups at AWS. “Expanding to the large and and unified US market is a major piece of the strategic roadmap for these startups as they look to widen their addressable market opportunites .”

Like every initiative at AWS, AWS Expand is driven by customer obsession. The program grew out of a series of interviews AWS conducted last year with startup founders and executives concerning the process of expanding a company beyond its domestic borders. The team at AWS learned that startups seeking to develop internationally face a number of challenges that can inhibit their growth potential—problems like a lack of strategy, planning, and roadmapping resources, or a need for operational support as they scale their physical presence in a new region. Most relevant to AWS Expand, though, is that they also encounter obstacles in their go-to-market efforts and in identifying and implementing the right technological solutions.

The AWS team knew it could already offer a robust response to startups’ tech needs through its Solutions Architects, but it also recognized an opportunity to provide go-to-market guidance and support by combing AWS’s vast investor and partner network with MATTER’s industry expertise.

Based in Chicago, MATTER has fostered a network of industry leaders and innovators working throughout the US healthcare sector. Its community features pharmaceutical manufacturers, insurance companies, hospital systems, educators, and more—all groups that are eager to engage with promising developments in healthcare and medical technology from overseas.

With its supportive community and comprehensive understanding of the industry, MATTER also provides emerging companies with essential education on the US market’s unique regulatory environment.

“The US healthcare industry is vastly different from other local healthcare markets. Its nuances and intricacies require a distinct knowledge base,” says Steven Collens, CEO of MATTER. “MATTER is ready to help these startups adopt a new lens that will be critical to their success in the US.”

Through this partnership, AWS and MATTER will advise on strategy and bring European startups and potential customers together, eliminating traditional barriers to entry for the US marketplace.

AWS also works closely with the venture capital firms that invest in these startups to ensure that they are well-positioned to thrive internationally. Saul notes that VC firms have a vested interest in seeing their startups succeed, but they recognize that international expansion is a major area in which these growing companies often need assistance. AWS collaborates with top VC firms in EMEA, relying on their knowledge to nominate the startups in their portfolios that are the most mature and most prepared to enter the US.

The advantages of being selected for this program are readily apparent to those involved. “AWS Expand presents a unique opportunity to advance the Sapien business with access to industry leading experts, extensive educational resources, and potential clients,” says Matthew Beatty, CEO of UK-based Sapien Health. “This initiative will help Sapien transform perioperative care for people across the US and in doing so help millions of people have a successful surgery.”

Dr. Philipp Brunnbauer, Co-Founder & CEO of MedEngine, based in Germany, notes that the EU’s new medical device regulations, vastly increased the need for additional resources to deploy consumer-focused products in the German market.

“We are looking for any kind of regulatory, legal or business development assistance to enter the US market, primarily focused on the East Coast,” he says.

AWS is focused on maximizing impact in the healthcare industry through this collaboration with MATTER. But the team recognizes the potential for assisting startups through similar partnerships in other industries in the future. By prioritizing a single vertical, AWS can dive deep into the specifics and learn what practices are most effective for firms making the leap to international expansion. From there, possibilities in other industries may open up. When they do, AWS will be ready to provide the support startups need to make the leap.

Learn more about how AWS is supporting Healthcare startups.

Video Series: The Startup Guide to AWS Services

AWS Editorial Team — Tue, 21 Jun 2022 17:46:16 +0000

As a founder, getting up and running on AWS, even knowing where to start can seem overwhelming. What services do you need? How do you build with best practices in mind? This series is your guide to getting started on AWS, from account setup and security, to choosing an operational model and database selection. Come explore the AWS cloud environment with us.

AWS has assembled a guide to help startup founders and developers identify products and services you need to get to MVP quickly. The Startup Guide to AWS Services series consists of seven short videos and blog posts featuring AWS Solutions Architects, with related links and training videos. Each presentation highlights the advantages to working in a cloud environment, with brief examples that pair services to use cases.

The series starts with security as the cornerstone of any startup. AWS works on a shared responsibility model, maintaining the security of the cloud, while the startup maintains security in the cloud. Building a Strong Security Foundation describes how to build security into a business, its operations, and production, from the start, while Building the Foundation for Your Cloud Environment covers managing permissions to control access.

Accelerating Your Development Journey and Rapidly Building and Delivering Your Application show how automation and analytics combine to develop your product—and startup—efficiently and securely. AWS recommends that all aspects of your business use cloud technologies to boost innovation, productivity, and security.

Automation is one hero of the AWS story. It makes speed possible with consistent, reliable processes. It also keeps a business working around the clock. Tracking analytics is another key feature. Data—securely managed—flows into the system 24 hours a day, seven days a week, and generates reports, turning rows of data into actionable insights. Alarms and alerts prompt the corresponding response, for example, to make and deliver more or less of a product or review pending financials.

To that end, accessibility is another core theme of the series. In the real world, someone might develop the idea for a winning app without having the skills to build, test, deploy, and update it. Even advanced technologies such as Artificial Intelligence (AI) and Machine Learning (ML) are packaged into AWS on a pay-as-you-go basis, making them accessible to those without prior experience. This way, startup founders and developers can focus on running their business, rather than building every element from the ground up. Advancing for Scale and Maturity identifies some uses of AI and ML in AWS products, while Choosing the Right Architecture and Choosing the Right Database cover how to choose and combine these crucial elements into a single, well-designed system.

A startup has many moving parts, each of which might call for different technologies. To build your company into a finely tuned growth engine, check out The Startup Guide to AWS Services video series. Then, start building!

The Sustainability Equation: Why Tech for Good Means Tech for the Future

AWS Editorial Team — Fri, 17 Jun 2022 15:02:18 +0000

The World Economic Forum (WEF) believes that technology is key to helping the world respond to global problems such as climate change. It’s technology that has allowed the world to redefine itself in the wake of a global pandemic, and that has given individuals and organizations the tools they need to thrive in complex times. We need startups with innovative mindsets to step forward with solutions that will enable the world to achieve the Sustainable Development Goals (SDGs), startups that put sustainability at the heart of their technology, which is precisely what Basecamp Research, BioSimulytics Limited, and Circular Way have done. These three companies have recently been recognized for their innovative approaches in the AWS Software Startups Awards. Receiving Gold, Silver and Bronze in the Sustainability Champion category respectively, these companies won because they’ve effectively leveraged technology for the betterment of society and the environment. Today, meet three software startups leading the sustainability movement in 2022.

Gold medalist Basecamp Research is leveraging tech to boost biodiversity

Basecamp Research is a young startup founded in 2020 that’s committed to the value and preservation of biodiversity. The company successfully completed its seed round in 2021—attracting $7 million from leading US and European venture capitalists. The ethos that drives this fresh-faced sustainability company is the belief that biotechnology is the first credible chance to wean the world off petrochemistry and to move towards a cleaner, biochemistry-based future that sustains biodiversity.

“In 2019, our team led the first ever fully off-grid environmental sequencing expedition to Iceland’s icecaps,” says Philipp Lorenz, Head of Data Science at Basecamp Research. “We were fascinated by the diversity of undiscovered species, genes and proteins in our planet’s remote and underexplored environments. This fascination led to us founding Basecamp Research.”

With just 17 employees, the company is already working to create a bridge between biodiversity and the bioeconomy, and promoting the protection of Earth’s wild places. Its goal is to identify the building blocks for a cleaner and more sustainable future. To achieve this goal, Basecamp Research is partnering with biodiverse nations to discover novel genes, enzymes, and organisms from diverse environments, and combining exploration and data science to build the world’s largest ethically sourced and fully contextualized Knowledge Graph about the planet’s undiscovered biodiversity. This Knowledge Graph uses the labels, annotations, and predictions generated by the London-based company’s machine learning pipeline to provide novelty, speed and improved performance to all sectors within the bioeconomy.

“By doing this, we provide biodiversity guardians with a fair and equitable way of sharing their environmental genomic data with biotechnology innovators,” says Lorenz. “It allows us to trace the benefit sharing agreement associated with every one of our products so we can ensure full traceability, and so we can use the benefits of our work to promote and protect the source biodiversity.”

This approach is vastly different from how the biodiversity and biotechnology sectors have communicated in the past and a step towards a more inclusive and collaborative future. As Basecamp Research focuses on the foundations for the planet’s future, it uses AWS FSx for Lustre, Amazon Elastic Block Store (EBS), and Amazon Elastic File System (EFS) for storage, and it has implemented Amazon EC2, Amazon CloudFront, and AWS Lambda, among other services, for its compute functionality.

“The most useful AWS cloud products for us have been AWS Step Functions and AWS Batch that we’ve used for the orchestration of our cloud production pipeline, BaseScan,” concludes Lorenz. “This pipeline was developed in-house and generates complex networks of annotations for our ethically sourced metagenome samples that we embed into our Knowledge Graph.”

Winning Gold in the AWS Awards has shone a light on the work done by this dynamic startup—work that really does leverage technology to power a more sustainable future. When asked what advice he would give to anyone wanting to start a tech-for-good enterprise, Lorenz said: “Have clarity in your vision because conviction and communication are absolutely key.”

BioSimulytics Limited won Silver for driving forward drug development

BioSimulytics Limited is a spin-off company from University College Dublin (UCD). Founded in November 2019 thanks to substantial national scientific grant funding, the company has developed an AI-powered molecular simulation technology platform designed to dramatically improve the accuracy, speed, and cost effectiveness of new drug development.

“The company came about after several years of scientific research into solving a long-standing and fundamental problem in the field of material science—the ability to predict how molecules will form as a crystal structure in their solid state,” says Peter F. Doyle, CEO and Co-Founder of BioSimulytics Limited. “This solid state structure impacts on everything from the taste of chocolate to the color of paint to the efficacy of a new drug.”

The company was motivated to solve this problem, not just because the research would be scientifically ground-breaking, but because it can potentially transform new drug development. Currently, bringing new drug molecules to market typically requires between $2-3 billion and up to 12 years and, even then, there is a limited chance (<1%) of success.

“One of the complicating factors in the drug development process is polymorphism, which is the ability of a compound to exist in more than one stable crystalline structure, if it can even be crystallized at all,” says Doyle. “Drug molecules are complex compounds that can have many plausible structures, and different drug polymorphs can have different properties, such as solubility, toxicity and efficacy.”

It is vital that pharma companies fully characterize the fundamental structural makeup of drug molecules and have absolute certainty about identifying and reproducing the most desirable crystal structure of any new drug before bringing it to market and patient use. The problem is that the lead optimization phase of drug discovery and development that accounts for 25% of the overall cost requires the synthesis of thousands of molecules to find the right one. By applying AI and machine learning technologies for molecular simulation and crystal structure prediction, BioSimulytics can essentially digitize a significant amount of the work carried out today.

“This reduces the amount of experimentation work down to only the most value-adding tasks and dramatically shortens the odds while accelerating the time-frame associated with new drug development,” says Doyle. “This will help usher in a whole new era of precision medicine for tackling many chronic and hard-to-treat diseases and this is the quest that drives our growing team today.”

BioSimulytics deploys its software technology for molecular simulation on an AWS High Performance Computing (HPC) parallel cluster platform. This is essential for the company to offer its global clients a cloud-based solution where they can run simulations on candidate drug molecules on an as-needed basis without compromising on performance, speed or compute.

For any entrepreneur thinking of setting off down the same path, Doyle has this advice: “Clearly articulate the societal challenge you are addressing and how you’re uniquely doing so, and then stay super-focused on this mission while remaining flexible enough to adapt the ‘how’ as you grow.”

Bronze award winner Circular Way is advancing sustainability in the retail sector

Circular Way was founded in 2021 and is dedicated to making circularity a central part of every purchase. It is working towards building the world’s first fully circular eCommerce system driven by engaging customer experiences. The company is innovatively meeting the retail sector’s need to improve its sustainability and environmental footprint by focusing on the circular economy. The Circular Way solution makes resale a part of every retail transaction, allowing for customers to cash-in previous purchases when buying new items.

“Our team believes in the urgency to transition the world to a circular economy to help tackle the current climate crisis and to lead towards a more sustainable future for the planet,” says James Pamplin, CTO and Co-Founder of Circular Way. “Our mission is to demonstrate how circularity works in a fully closed-loop system, and how our technology is a key enabler for this. We were founded with this mission in mind and remain focused on solving these critical issues today.”

The company is in the early startup stage and has initial seed funding secured with a new funding round in the pipeline. It has already launched its pilot sell-back experience to two retail partners on the Shopify platform, allowing for customers to seamlessly sell any item they’ve purchased at that retailer when they buy a new product.

“The pilot results have shown great customer interest with up to 67% higher email campaign open rates and up to 80% higher email click rates,” says Pamplin. “A/B test results have also shown an improvement to checkout conversion rates for customers that have been offered the ability to sell back their items.”

AWS is a key technical enabler for Circular Way’s platform, providing the backbone of its eCommerce integration through Amazon API Gateway, AWS Lambda, Amazon DynamoDB, Amazon S3, AWS Glue, and Amazon Athena, among others.

“We focus on using several AWS services for our core platform—their strength lies in enabling us to work seamlessly across multiple services,” concludes Pamplin.

When asked what single piece of advice he would give to anyone about to step into tech-for-good, Pamplin says: “People are key, so ensure that the people you work with, as well as your business partners, align with your company values and mission as this helps create effective and productive working relationships as the business grows.”

The AWS Software Startups Awards recognizes innovative startups and entrepreneurs across several key categories, including Rising Star, Go-To-Market Innovator, Founder of the Year and Rocketship. Become a member of the AWS Startup Loft for news on opportunities like this and more.

A Startup’s Guide to AWS Services Series 7: Choosing a Database

AWS Editorial Team — Wed, 15 Jun 2022 13:52:33 +0000

As a founder, getting up and running on AWS, and even knowing where to start can seem overwhelming. What services do you need? How do you build with best practices in mind? This series is your guide to getting started on AWS, from account setup and security, to choosing an operational model and database selection. Come explore the AWS cloud environment with us.

Choosing the right database

Click image to watch video

Data needs to live somewhere. Whether for ETL solutions, data cataloging, or advanced analytics, data needs to be stored safely and accessibly. In Choosing the Right Database, the seventh installment of The Startup Guide to AWS Services free video series, AWS Startup Solutions Architect Zoish Pithawala discusses how to form an ecosystem where data is mobile, agile, and secure.

“Many startups find that relational and key-value data stores are ideal for many use cases,” Pithawala points out. “But there are times when a more purpose-built database would offer benefits, such as performance and scale.” She divides databases into eight categories based on their structure and uses:

Relational
Key-value
Document
In-memory
Graph
Time-series
Ledger
Wide-column

The categories are matched with relevant AWS services such as Amazon Aurora, Amazon DynamoDB and Amazon DocumentDB (with MongoDB compatibility), Amazon Neptune, and Amazon Keyspaces (for Apache Cassandra). Use cases are given for each category, ranging from finance, shopping carts, content management, real-time analytics, and fraud detection to IoT applications, supply chains, and healthcare. Often, multiple database types should be combined in one system.

A startup producing an e-commerce application might have a product catalog requiring low latency. That would suit a key-value database such as DynamoDB. Adding a bidding component could call for Amazon ElastiCache for in-memory data storage. Then, to capture user clickstream data from the application, Pithawala recommends a time-series database, such as Amazon Timestream, which collects, stores, and sequences data by time.

“Make sure you don’t use a Swiss-Army-knife approach,” warns Pithawala. “Choose the right database for each individual component that you build. As you scale and find that your database becomes your bottleneck, you can start optimizing your database type for different use cases.”

For many companies, choosing a database really means choosing the right combination of databases. Identifying the needs of each component of the business—and matching them to suitable databases—is the first step. Then, over time, you can be flexible in making changes as your business grows.

Two Awesome Days at AWS Summit Atlanta

AWS Editorial Team — Fri, 10 Jun 2022 11:13:44 +0000

Post and video by Chalaire Miller, Web and Social Content Manager, AWS Startups

To all those who attended the AWS Summit in Atlanta, what a great time we had! For those who were unable to attend, we shared two amazing days with founders, developers, engineers, IT executives, and technical decision-makers from all over the country. Our keynote speaker, Miriam McLemore, Director of Enterprise Strategy and Evangelism at AWS, shared some amazing insights and innovations within the AWS platform aimed at helping startups accelerate to MVP. We heard from experts in AI/ML, Analytics, Business Intelligence, IoT, Databases, and more, who connected and networked with industry peers to help them take advantage of the latest AWS solutions.

Here are just a few reasons to attend an AWS Summit:

Connect with and learn directly from AWS experts.
Network and interact with other startup founders.
See the latest and greatest AWS innovations.
Be inspired by industry experts and leaders of AWS solutions.

Interested in attending future AWS Summits? Visit aws.amazon.com/events/summits to register for one of our global events today.

A Startup’s Guide to AWS Services Series 6: Choosing an Architecture

AWS Editorial Team — Thu, 09 Jun 2022 12:02:49 +0000

As a founder building to MVP on AWS, knowing where to begin can feel overwhelming. What services do you need? How do you start building with best practices in mind? This series is your guide to getting started on AWS, from account setup and security best practices, to choosing an operational model and database selection. Come explore the AWS cloud environment with us.

Choosing the Right Architecture

Click image to watch video

Compute options: The first decision made by startup founders and developers radiates out to affect all aspects of the business. Rather than imposing limits on what can be designed, built, or deployed, Principal Startup Solutions Architect Igor Geyfman explains that a single system can incorporate different compute options. On this model, each aspect of the system is designed with a view toward “performance and scalability, the level of maintenance required, and the cost.”

Choosing the Right Architecture, the sixth installment of The Startup Guide to AWS Services video series, covers AWS compute models and the advantages of serverless architecture. While outlining the various operational models, Geyfman identifies which tasks AWS manages and which are up to the startup. Serverless architecture is mainstream now, Geyfman points out. The AWS models are arranged on a scale from those requiring the least user involvement to those requiring the most, providing more flexibility. As your startup grows, its needs might change. “We always suggest you start with AWS Lambda, because you’ll manage as little as possible,” says Geyfman.

AWS Lambda offers serverless functions, data source integration, physical hardware, software, networking, and more. Here, the startup writes the application code, and AWS manages the rest, potentially speeding up the time to market and supporting innovation.

If the Lambda service seems too limited, the next level would be AWS Fargate, a serverless container service. While AWS orchestrates the containers, among other tasks, the startup’s responsibilities include data source integration, security, and management.

As your startup progresses, you may want to assume more control, taking on additional responsibilities and direct costs. If so, Geyfman suggests using Amazon Elastic Container Service or Amazon Elastic Kubernetes Service, which are instances of container management as a service.

For even more control, startups can run an application directly on Amazon Elastic Cloud Computer (Amazon EC2), which offers infrastructure as a service, with the highest level of management by the startup. Here, a startup would deploy and run an application directly on EC2 compute instances. AWS is responsible for physical hardware, software, networking, and facilities, while the startup is in charge of everything else: from writing the application code to scaling and management tasks.

Along with the range of operational models, Geyfman explains that “there are serverless services at all layers of the stack.” Starting with the serverless compute layer, he moves on to discuss serverless data stores, such as Amazon Simple Storage Service, and serverless integration services, from Amazon EventBridge to AWS AppSync. These, he explains, allow you to “connect your microservices and applications, as well as integrate with external systems.”

The ideal architecture for your startup may change substantially over time. AWS helps you to hit the ground running, then adapt as needed, taking on more hands-on management as you progress through the range of serverless architecture options.

Watch the full episode to learn more.

AWS Announces 25 Startups Selected for Inaugural AWS Impact Accelerator for Black Founders

AWS Editorial Team — Wed, 08 Jun 2022 13:08:31 +0000

More than 15 years after AWS changed the game for startups by giving them access to the same technology as the world’s largest enterprises, we continue to level the playing field so that founders can develop their ideas and build lasting businesses regardless of gender, sexual orientation, ethnicity, or race. In April, we announced a 3-year, $30 million commitment to the AWS Impact Accelerator, a series of programs is designed to help high-potential, pre-seed startups led by underrepresented founders succeed. The first of these programs, the AWS Impact Accelerator for Black Founders, invited applications from U.S.-based startups to an overwhelming response, with applications pouring in from across the country.

Today, we’re excited to announce the startups that will make up this inaugural cohort. A diverse committee of AWS startup experts selected these 25 participants based on the strength of their idea and technical readiness, in addition to a series of interviews. They represent an inspiring, eclectic mix of industries, use cases, business models, and geographies:

Arizona

PeerCapsule – PeerCapsule, based in Phoenix and New York City, is a first-of-its-kind collaboration network for verified healthcare professionals to exchange clinical insight, find research mentors, and network with experts.

California

Oben Health – Oben Health, based in San Francisco, is developing a digital therapeutic to safely and sustainably reverse heart disease without medication or surgery, helping to minimize an epidemic that disproportionately kills Black and brown patients.

CModel – Executives use Vallejo-based CModel’s real-time revenue analysis platform to evaluate their revenue strategy with concise, actionable, and accurate intelligence.

HubbleIQ – Bay Area-based HubbleIQ’s software provides automated remote tech support that is easy, data-driven, and deployable in minutes, so that businesses of any size can operate with more agility and efficiency.

Seed at the Table – Compton-based Seed at The Table is a mission-driven crowdfunding platform committed to connecting diverse entrepreneurs with accredited and non-accredited investors looking to obtain equity in BIPOC-led companies.

SevenStar – Designed to help bridge the trust gap between law enforcement agencies and citizens, Irvine-based SevenStar’s cloud-based app enables critical, real-time information sharing for safer and more accountable operations.

Delaware

Flywallet – Recognizing that 7 out of 10 travelers go into debt for a trip, Flywallet created a gamified savings account designed to make travel more affordable.

Rush Roto – Rush Roto, based in Dover, uses AI to help small businesses create professional product photos without the overhead of typical studio photography.

Florida

Imanyco – West Palm Beach and St. Louis-based Imanyco provides a real time transcription app that helps the deaf and hard of hearing quickly and reliably understand and participate in group conversations.

Georgia

Chezie – Based in Atlanta, Chezie helps companies retain diverse talent with an all-in-one solution for building and managing impactful employee resource groups.

EyeGage – EyeGage, also based in Atlanta, increases the speed and accuracy of drug detection with computer vision analysis of the eye, ensuring greater safety in high-risk environments.

The Goji Shop – Atlanta-based Goji is using conversational AI to build a personalized customer service layer that analyzes products and helps shoppers make smarter purchasing decisions, reinventing online shopping.

Reflekt Me – Reflekt Me, based in Marietta, helps brands reach 100% of their customers with a hyper-personalization platform that matches the characteristics of each online shopper with existing perfect product details page (PDP) images.

Illinois

The Crypto Mom – Based in Chicago, The Crypto Mom balances the male-dominated crypto space by providing a path for women to discuss, purchase, and give crypto.

CulturePi lot – CulturePilot, also based in Chicago, offers a single platform that frees museums and exhibition spaces to create Computer Vision-powered tours for visitors without needing to invest in their own custom equipment.

Maryland

CyDeploy – Based in Baltimore, CyDeploy provides an intelligent, automated approach to functional testing of IT and Internet of Things (IoT), helping to prevent the 80% of security breaches that occur due to misconfigured or unpatched systems.

New York

CareCopilot – With an elder care crisis approaching, New York City-based CareCopilot provides an app that guides loved ones through all of the steps necessary to care for aging relatives.

Gyvlink – This Albany-based SaaS platform reduces donation waste, donor churn, staff costs, and warehouse charges by helping organizations rapidly share and match requests for in-kind donations.

Oarbt – Oarbt, based in New York City, is a no-code, AI-powered SaaS design solution that enables businesses to launch and gather actionable insights from interactive 3D stores.

Recut – Recut, based in New York, creates compelling video advertising experiences by splicing existing actor-based video content with AI-generated photorealistic video and audio.

Rhode Island

Solvent – Solvent, based in Providence and New York City, provides affordable personal finance and banking tools for people who have been impacted by systemic issues like financial exclusion, mass incarceration, and the wealth gap.

Tennessee

Nodat – Nodat, based in Nashville, is a geo-targeting mobile marketing platform that closes the local marketing gap by connecting businesses to the right consumer in the right service area.

Texas

ChurchSpace – Coined the “Airbnb for churches,” Houston-based ChurchSpace is the only marketplace designed specifically to help churches turn their unused space into on-demand event, meeting, and commercial kitchen space.

Heroshe – Heroshe, based in Houston, provides an ecommerce and logistics enablement platform that allows social sellers in Africa to grow their cross-border business capabilities.

Moyae – Moyae, based in Austin, provides a software suite that enables ophthalmologists and optometrists to create compliant records, maintain relationships with patients, and manage billing while also improving the speed of feedback between researchers and doctors.

Each of these startups will receive up to $225,000 in cash and AWS Activate credits, an extensive and individually-curated training curriculum, mentoring and technical guidance, introductions to Amazon leaders and teams, networking opportunities with potential investors, and ongoing advisory support. They will also have the opportunity to learn from organizations AWS works closely with such as Black Women Talk Tech (BWTT), digitalundivided, and Visible Hands, as well as hear from successful founders like Dave Salvant, President and Co-founder of Squire, a startup that offers a technology platform for barbershops.

AWS will continue to deliver guidance and resources after the 8-week program ends, and these founders will have access to a virtual community, alumni events, the curriculum (including future content), plus the opportunity to tap into ongoing support from program mentors and experts.

The first AWS Impact Accelerator for Women Founders will begin accepting applications in July, with the AWS Impact Accelerator for LGBTQIA+ Founders and AWS Impact Accelerator for Latino Founders to follow in 2023. Become a member of the AWS Virtual Loft to be notified when applications are open.

A Startup’s Guide to AWS Services Series 5: Analytics and Automation – Superhighways to Scale

AWS Editorial Team — Wed, 01 Jun 2022 10:43:34 +0000

Advancing for scale and maturity

Click image to watch video

Everything’s going well. The product is out there, and customers are using it. But suddenly, you find yourself on unfamiliar ground: you need to keep building and deploying updated versions of the product in response to customer needs. And you need analytics flowing in continuously to identify what those needs may be.

This is a crucial intersection for startups, when time and resources are often stretched as far as they can go, and various branches of the infrastructure need to communicate. The Startup Guide to AWS Services free video series can help startup founders and developers bridge the gap, obtaining the tools they need without having to build them from scratch. Automation—using machine learning—can put startups on a growth superhighway.

In Advancing for Scale and Maturity, AWS Startup Solutions Architect Zoish Pithawala describes a two-pronged approach for startups scaling an application. “At this stage, it’s essential to review the useful data your application is collecting and apply those insights to improve the experience for customers.” To stay nimble for continuous adaptation, you need detailed analytics, allowing you to anticipate customers’ needs. Then, Pithawala explains, you need automation and integration to reduce the workload for the staff. Taken together, these twin engines—analytics and automation—keep information flowing in and incorporating into a product that is effective and up to date.

AWS Artificial Intelligence and Machine Learning Services (AI/ML) are powerful tools, which put machine learning into the hands of every developer, allowing them to add intelligence to any application without needing ML skills. Your individual use case will help identify which service to implement for your product, such as personalization or fraud detection.

Pithawala calls AWS Analytics “the fastest way for me to turn my rows of data into answers.” You can implement a data strategy and make data-driven decisions throughout your business. Meanwhile, the AWS Application Integration suite can be applied to microservices, distributed systems, and serverless applications to allow for communication between decoupled components. This can help with API management, messaging, and workflows, among other tasks.

Technical founders may already be proficient in these various fields. If not, Pithawala recommends working with specialists. For ML solutions, for example, startups can enlist the aid of data scientists to help them get started.

Data gathered can be an invaluable resource, as long as it is analyzed in a timely manner so you can act on it. Then, automation helps implement the changes needed in ways that are manageable for your busy staff. With these two service solutions together, AWS makes advanced technologies accessible to all, while keeping your information flowing and up to date.

June in San Francisco: Lunch & Learns, Workshops at the AWS Startup Loft

AWS Editorial Team — Tue, 31 May 2022 15:59:00 +0000

In April, we announced the re-opening of the San Francisco AWS Startup Loft, an innovative space open to all startups building on AWS, and a free benefit for existing AWS Activate members.

The San Francisco AWS Startup Loft is a hub where founders, builders, and their teams gain access to the expertise they need to build, launch, and scale their startup. This month, you’ll find six exclusive in-person events for startups looking to optimize your product for better product-market fit and understand how to build to MVP.

To register, simply visit the Loft page, and click “sign up” to create a profile. From there you’ll have access to free Loft co-working, events, networking, and mentorship. Current Loft members only need to sign in.

June 8: Lunch & Learn: This Isn’t That MVP Session

Time: 10:30AM – 1:00PM (Pacific Daylight Time)
Speaker: Mike Apted, Startup CTO Engagement Lead

You’re eager to build your MVP: something viable you can take to an investor. And while AWS strives to help founders get started, this isn’t THAT MVP session. Instead, this is an MVP workshop that goes beyond the tools, to focus on how to get started fast and validate your assumptions with the least amount of work (and paying customers).

Designed by a former founder-turned-AWS Startup Solutions Architect, this session dives into managed services, optionality, risk reduction, how to test for free with AWS credits, and how to iterate based on user behavior. We’ll also review good (and not so good) MVP examples. Come with questions, leave with answers, ready to build an investor-worthy MVP. Sign in / sign up to RSVP.

June 9: Lunch & Learn: Founder-Led Sales

Time: 10:30AM – 1:00PM (Pacific Daylight Time)
Speaker: Mark Birch, Principal Startup Advocate

Are you a startup founder that has never sold a product before, and you need tactical strategies and steps to do so? In this session, we’ll go over the 12 steps to start selling with confidence and consistency. You’ll learn how to build a Motivation Matrix to unpack customer problems and the formula for the right customer messaging. You’ll then develop initial lead lists and use templates for prospecting. We’ll end with deal qualification, running a sales meeting, and closing the deal. This session is appropriate for early-stage B2B/B2B2C startups looking to uplevel and systematize their sales process for scale. Sign in / sign up to RSVP.

June 14: Lunch & Learn: From 1 to 200 – Choosing the Right Solutions for Your Startup at the Right Time

Time: 12:00PM – 3:00PM (Pacific Daylight Time)
Speaker: Igor Geyfman, Principal Startup SA

You have a vision for how you want your product to perform. Now, you just need to put the right tools in place to make it happen. But with over 200 AWS services, how do you decide what’s best for your startup?

Whether you’re looking to go serverless, build containers, or want to get something up and running quickly to validate assumptions, AWS Startup Solution Architect Igor Geyfman will break down the key services startups need to consider on Day One. He’ll also help you look ahead at the essential services to add as you scale. This session is well-suited for early-stage technical founders and startup cloud architects. Sign in / sign up to RSVP.

June 16: Lunch & Learn: From 1 to 10 Million Users

Time: 10:30AM – 1:00PM (Pacific Daylight Time)
Speaker: Itzik Paz, Sr. Startup Solution Architect

Cloud computing gives you a number of advantages, such as the ability to scale your web application or website on demand. But if you have a new web application and want to use cloud computing, where do you start? Join us in this session to learn best practices for scaling your resources from one to millions of users. We show you how to best combine different AWS services to optimize systems and costs and how to make smarter decisions for architecting your application. This session is great for both B2C and B2B companies who are looking for a relevant and practical framework for thinking about scale. Sign in / sign up to RSVP.

June 21: AWS Lightsail Workshop: Spinning Up Your Web App Fast

Time: 9:30AM – 1:00PM (Pacific Daylight Time)
Speaker: Kunal Batra, Developer Advocate Manager & Sisira Narayana, Principal Product Manager

For startup developers who need to build websites or web applications, Amazon Lightsail is the easiest way to get started. In this half-day workshop, learn how to create a WordPress website in minutes without deep technical knowledge. You will learn how to collect site visitor information, share job descriptions, and utilize automation to keep your customers updated on the latest news. We’ll also walk you through how to configure custom domains, modify the website copy/theme, and integrate Amazon SES for email communication with users/customers. You’ll also deploy a LAMP stack application with a multi-channel communication service to communicate with your users through SMS, email and more. Sign in / sign up to RSVP.

Please allot approximately 2.5 hours for the full workshop.

June 23: AWS Amplify Workshop: Get Your MVP to Market Faster

Time: 1:30PM – 7:30PM (Pacific Daylight Time)
Speaker: Ritik Khatwani, Startup Solutions Architect & Jessie VanderVeen, Principal Product Manager

Build and and ship complete apps in hours instead of weeks and launch your MVP with AWS Amplify Studio! In this workshop, jointly presented by Figma — a web-based design and prototyping tool — and AWS Amplify Studio, you’ll learn how to serve your front and back-end development needs. You can build an app backend, create custom user interface (UI) components, enable designers to customize components in Figma, which can be imported into Studio, and connect a UI to an app backend with minimal coding. Developing a full-stack app on AWS just got easier, and this session shows you how. Sign in / sign up to RSVP.

Please allot approximately 2.5 hours for the full workshop.

We look forward to seeing you at an AWS Startup Loft event soon!

ICYMI: A Startup’s Guide to AWS Services, Experiences from AWS Summit San Francisco That Will Give You FOMO, and More

AWS Editorial Team — Fri, 27 May 2022 11:39:03 +0000

Announcements

Seven Experiences from AWS Summit San Francisco That Will Give You FOMO

Thousands of attendees and dozens of vendors gathered at AWS Summit San Francisco, part of a series of free, global, in-person events dedicated to spreading skills and knowledge around AWS’s cloud offerings. Here are seven of those attendee experiences that are guaranteed to give you enough FOMO to immediately register for an AWS Summit near you!

Learn How to Grow Your Startup with Machine Learning on Twitch

While Machine Learning can get quite complex, you don’t need a team of expensive Data Scientists and ML Engineers to gain real value from it. Check out our upcoming Twitch series for hands-on training with our AWS ML experts, and work through a variety of typical startup use cases from generating personalized customer recommendations to improving marketing efficiency.

Resources

Day One Recommendations to Secure Your Account and Workload with AWS Startup Security Baseline

Security is always the top priority for AWS, but let’s face it – it’s not always the top priority for a founder just trying to get your idea for a new company off the ground. That’s why we’re pleased to announce the launch of the AWS Startup Security Baseline (AWS SSB), a new guide that describes the set of controls we recommend all startups implement as a foundation during their initial stages of development and operation.

SERIES: A Startup’s Guide to AWS Services

Series host: AWS Startup Solutions Architect Zoish Pithawala

As a founder building to MVP on AWS, knowing where to begin can feel overwhelming. What services do you need? How do you start building with best practices in mind? This month, we launched the Startup’s Guide to AWS Services Series, your guide to getting started on AWS, from account setup and security best practices, to choosing an operational model and database selection. Stayed tuned for the second half of the series in June!

A Startup’s Guide to AWS Services Series 1: Making Security the Cornerstone

You’ve developed a winning concept, gathered a team, and mapped out your startup’s overall design. Now, as those ideas take shape, security solutions should be built right in.

A Startup’s Guide to AWS Services Series 2: Innovation And Control – Having It All

A Startup’s Guide to AWS Services Series 3: Speeding Toward an MVP

Series host: AWS Principal Startup Solutions Architect Igor Geyfman

Running a business is hard work, with a lot of moving parts. Startup founders have had to build, deploy, secure, scale, manage, and monitor their software products all while controlling operational costs. This installment of the video series matches services to situations to streamline the development process to develop and deploy an MVP fast.

A Startup’s Guide to AWS Services Series 4: Building Faster in the Cloud

Build. Test. Deploy. Troubleshoot. Then do it all again. For a startup trying to build and deploy an application, the process can seem endless. Luckily, automation and cloud-based resources can streamline the process.

Startup Spotlight

Securing Digital Interactions Within the Pharmaceutical Supply Chain With Spherity

Spherity is a German software provider bringing secure and decentralized identity management solutions to enterprises, machines, products, data, and even algorithms.

A Startup’s Guide to AWS Services Series 4: Building Faster in the Cloud

AWS Editorial Team — Wed, 25 May 2022 14:19:54 +0000

Rapidly building and delivering your application

Click image to watch video

Build. Test. Deploy. Troubleshoot. Then, do it all again—for every single update. For a startup trying to build and deploy an application, the process can seem endless, with time to market an ever-receding horizon.

Luckily, automation and cloud-based resources can streamline the process. In Rapidly Building and Delivering Your Application, the fourth installment of The Startup Guide to AWS Services video series, Principal Startup Solutions Architect Igor Geyfman identifies prefabricated code structures that streamline development. These can be combined with “templates for quick deployment and services to connect them to your broader build,” Geyfman explains. The startup decides when and where to use AWS technologies and services without upfront fees or a long-term contract.

First, startups need to build and manage an infrastructure. To do this effectively, Geyfman recommends separating development, staging, and production into different environments. The AWS Infrastructure as Code approach offers a consistent and repeatable infrastructure deployment. It has two routes to setting up the infrastructure: AWS CloudFormation or AWS Cloud Development Kit.

Next, users can build, test, and deploy the application automatically. AWS Code Pipeline automates the software release process, allowing for fast and reliable updates. This flexible delivery pipeline automatically runs the build–test–deploy process in response to code changes.

AWS Code Build is a continuous integration service for building, testing, and packaging. Users can compile source code, run tests, and produce deployable software packages. Geyfman suggests also using the AWS CodeDeploy service. By automating deployment to a variety of compute services, CodeDeploy streamlines the release of new features.

Once an application has been deployed, its performance needs to be monitored for what Geyfman calls “unexpected behavior in the environment.” Amazon CloudWatch is a service that monitors the system, while also providing data and analytics about the application for “visibility into how it’s performing.” Users can also set alarms to respond in a timely manner to issues that arise.

Prefabricated code structures, effective templates, and a strong mix of services can help your business progress by leaps and bounds. Put together, AWS helps startups grow and scale, with security oversight and analytics built right in.

A Startup’s Guide to AWS Services Series 3: Speeding Toward an MVP

AWS Editorial Team — Thu, 19 May 2022 11:30:00 +0000

Accelerating Your Development

Click image to watch video

Running a business is hard work with a lot of moving parts. On top of that, in the traditional model, startup founders have had to build, deploy, secure, scale, manage, and monitor their software products all while controlling operational costs. Fortunately, times have changed. Today, you can leverage an array of technologies to develop and deploy an MVP fast, so your business can grow and scale. This installment of The Startup Guide to AWS Services free video series matches services to situations to streamline the development process.

In Accelerating Your Development Journey, AWS Startup Solutions Architect Zoish Pithawala shows you how to use specific AWS services to shape a startup’s digital architecture. “Wouldn’t it be great,” asks Pithawala, “if there was a way to just focus on the application code, while the undifferentiated plumbing was created and managed automatically?”

AWS Amplify is a service that helps developers build serverless mobile and web apps that scale. It’s flexible, with features such as hosting, authentication, API, and data stores that can be used individually or in combination, or added over time, as a startup grows. Amplify supports various web frameworks and mobile platforms.

If you’re planning to use containers, you can create the source code or a container image, and then use AWS App Runner to take it from there. The service automatically builds the app, scales the infrastructure needed, and load-balances traffic with end-to-end encryption, among other functions. Even without prior infrastructure experience, Pithawala explains, startups can “quickly deploy containerized web applications and APIs to scale.”

AWS Elastic Beanstalk helps startups build more traditional compute instance-based backend services. You can use Elastic Beanstalk with applications developed using various programming languages and platforms, such as Java, Python, and others. At some point, startups might find they need a virtual private server (VPS) to build an application or website. Amazon Lightsail provides a VPS that is easy to use for a simpler workload, paid for on a monthly plan.

Finally, if none of those options fits the bill, you can explore the AWS Marketplace, which contains software listings from independent software vendors. This way, you can add to your startup’s technical skill set by connecting with someone who can help you build a product using AWS services and best practices.

Development can be streamlined. By automating systems and taking advantage of a strong combination of services, you can focus on running the business and helping it grow.

Learn How to Grow Your Startup with Machine Learning on Twitch

AWS Editorial Team — Tue, 17 May 2022 15:04:02 +0000

Done right, Machine Learning (ML) can be a major differentiator for your startup. While ML can get quite complex, you actually don’t need a team of expensive Data Scientists and ML Engineers to gain real value from ML – like improving customer acquisition or providing personalized recommendations for customers.

Ready to learn more? Then check out our upcoming Twitch training series, Let’s ship it – with AWS! ML Edition to learn how to easily get started with ML. This live, interactive training with AWS Machine Learning experts Aaron Hunter and Fred Graichen will begin on June 2nd and run through July 21st, every Thursday from 4pm – 5:30pm PT. Each session will feature a hands-on ML use case, and we’ll be answering your questions live. Designed for Twitch with startups in mind, this new training series marks the newest Twitch series from AWS Training and Certification.

“On Twitch, you can learn AWS from AWS Experts, which allows interactive collaboration through demos and discussion. As a learner you can ask questions in the chat and have real time impact on the training covered by the host. You can learn as part of a community with an average of 300 learners joining each episode LIVE on Twitch, who help each other answer questions and learn. I love being able to work directly with learners to share knowledge and information, while having fun!”- Aaron Hunter, AWS Senior Technical Trainer and Twitch host

Why you should join

Just like AWS’ ML training resources, startup developers can leverage ML from day one—even without a Data Scientist or ML background. In this upcoming Twitch training, you will learn the basic concepts and processes involved in Machine Learning with Amazon SageMaker, as well as how to implement them yourself. If you’re a developer working in a lean startup environment, this training is the perfect place to start—even if you have limited access to data or resources.

Throughout the training, you’ll get hands-on with our AWS ML experts and work through a variety of typical startup use cases from generating personalized customer recommendations to improving marketing efficiency. As we work through the examples together, you can lean on the instructors and AWS experts in the chat if you get stuck or have any questions. If you want to follow along in your own AWS environment, you can create an AWS account and build with us using the Free Tier.

How to sign up

This Twitch series is available to anyone as part of our commitment to train 29 million learners globally by 2025. You can sign up to be notified when each new training is about to air live on Twitch here. If you can’t join live, sign up to receive replay links in your inbox, so you can watch the on-demand episodes after they air. You’ll also get the resources used in each episode, and additional tools to help you on your Machine Learning journey. See you on June 2 at 4pm PT!

Day One Recommendations to Secure Your Account and Workload with AWS Startup Security Baseline

AWS Editorial Team — Fri, 13 May 2022 12:12:10 +0000

Security is always the top priority for AWS, but let’s face it – it’s not always the top priority for a founder just trying to get your idea for a new company off the ground. Your goal is to get a prototype of Your product or service in front of customers and investors as quickly as possible to test for market fit before the money runs out! It can also be hard for founders and engineers to know just what security controls you should put in place during these formative moments, let alone how to implement them on AWS. As a result, a lot of really critical controls can easily be skipped or overlooked, which can spell bad news for the startup if it leads to a security incident.

That’s why we’re pleased to announce the launch of the AWS Startup Security Baseline (AWS SSB), a new guide that describes the set of controls we recommend all startups implement as a foundation during their initial stages of development and operation. It focuses on two main areas – securing your AWS account, and securing your workload – and provides all the step-by-step instructions needed for each control with links to the relevant documentation. These controls also align to the best practices of the Security pillar of the Well-Architected Framework, establishing a strong foundation from which you can evolve your security posture as your company grows.

Secure your account

The section on securing your account focuses on good IAM practices, preventing accidental misconfigurations, and setting up monitoring for threats and other risks. It includes well-known controls such as enabling MFA on Root, setting up IAM password policies, and enabling CloudTrail delivery to S3, as well as additional controls such as setting up AWS Budget alerts, enabling Amazon GuardDuty, and monitoring AWS Trusted Advisor.

Secure your workload

The section on securing your workload emphasizes managing application secrets and scope of access, minimizing access routes to private resources, and using encryption to protect data in transit and at rest. It includes controls such as using resource-based policies, encrypting Amazon RDS and Amazon EBS volumes, using Amazon VPC endpoints to privatize traffic flows, and using AWS Systems Manager for remote sessions.

AWS SSB in action

Startups are already sharing stories about the positive impact they’ve seen from implementing the controls of the AWS SSB.

“Following the AWS Startup Security Baseline, we enabled Amazon GuardDuty in our account, and it detected outside IPs trying to access our servers via Secure Shell (SSH). We removed port 22 (SSH) from all of our security groups and switched to using AWS Systems Manager to access our Amazon EC2 instances now. With these changes, we no longer see attempts being made by these external actors. This is just the first big benefit we received from following the recommendations in the guide!”
– Bob Lee III, CTO of ConnectCareHero

All of the controls in the AWS SSB are able to be implemented quickly and easily and do not require any security expertise. Businesses operating at later stages and higher scale can still derive a lot of value comparing their current controls to the ones provided in the baseline to identify any gaps.

Next Steps

For founders just getting started, take a few moments to work through the guide and ensure your AWS account is secured. Review your current workload security and see which controls need to be applied as well. Use AWS Trusted Advisor to assess your current security posture and resolve any high-risk items using the recommendations of the AWS SSB.

For anyone that has security-related questions or would like to speak to someone about their security needs, you can find out more on our AWS Cloud Security page. Have fun and stay secure while building quickly!

Seven Experiences from AWS Summit San Francisco That Will Give You FOMO

AWS Editorial Team — Thu, 12 May 2022 12:33:41 +0000

Post by Chalaire Miller, Web and Social Content Manager, AWS Startups

Thousands of attendees and dozens of vendors gathered at the George R. Moscone Convention Center on April 20 and 21 for AWS Summit San Francisco, part of a series of free, global, in-person events dedicated to spreading skills and knowledge around AWS’s cloud offerings.

You probably have innovation or data decisions you should be deploying right now. but, instead of doing that, here are seven experiences from #AWSSummit San Francisco that are guaranteed to give you enough FOMO to immediately register for an AWS Summit near you!

1. Attendees were inspired by the enthusiasm brought by Swami Sivasubramanian—Amazon Web Services’ Vice President of Data, Analytics, and Machine Learning Services—in his keynote address. They also benefited from Sivasubramanian’s insight into how customers in the Americas and across the globe are using AWS to transform their businesses.

2. One of the best reasons to attend an in-person event? You’ll get to hear exciting live announcements before anyone else! AWS Summit attendees were the first to learn about the AWS Impact Accelerator, a $30 million investment designed to support Black, Latino, Women, and LGBTQIA+ tech founders. Learn more about this exciting program!

3. Still on the fence about attending a live event? Let us convince you with three words: AWS DeepRacer League. Attendees had the opportunity to test their driving skills against the clock with the new AWS DeepRacer Arcade collaboration with F1. Discover if your #MachineLearning skills are competition-ready by going for a spin at an AWS Summit near you.

4. There’s no better way to get psyched about growing your company than learning about crucial topics from experts IRL. At the packed AWS Startups booth, a rapt audience hung onto Startup Solutions Architect Jillian Forde’s every word, as she shared her expertise on cost optimization and managing cloud costs for a growing business.

5. You had to be there to see it: Visitors at the Startup Summit Lofts watched a video featuring hip-hop legend and record executive Jim Jones. In it, Jones discusses how he started Quarantine Studios, a multifaceted platform that allows for virtual, real-time recording, with the help of AWS’s cloud technology. Watch the full video here.

6. AWS on Air is just one of many AWS Summit highlights. Special guests Senior Field Marketing Manager Betty Chung and Senior Startup Program Manager Derek Pham joined AWS on Air to discuss the reopening of the San Francisco AWS Startup Loft. The AWS Startup Loft welcomes innovators, entrepreneurs, builders, and visionaries alike to a space where anything is possible. Bring your vision—we’ll provide the resources that help make it real. Sign up today!

7. Virtual events have their benefits, but one thing they don’t have is the opportunity for in-person collisions—as these lucky attendees who ran into AWS Chief Evangelist Jeff Barr can attest! Conference goers had the opportunity to mix, mingle, and share ideas at an hour-long networking reception. Attend an in-person AWS Summit and you might just meet your next business partner—or, at the very least, walk away with some new inspiration.

Still not convinced that attending a live AWS Summit event is for you? Learn more about how AWS Summits are bringing technologists together to connect, collaborate, and learn about AWS.

A Startup’s Guide to AWS Services Series 2: Innovation And Control – Having It All

AWS Editorial Team — Wed, 11 May 2022 10:59:24 +0000

Building the foundation for your cloud environment

Click image to watch video

Founders don’t build startups in a vacuum. Starting up requires a collaboration of people and services, from product development to IT and compliance. You may find yourself pulled in many directions, and soon, you’ll have to face facts: you need to delegate. But doing so creates security risks.

“Anybody with access to my accounts can intentionally or unintentionally make irreversible changes in the account,” warns AWS Principal Startup Solutions Architect Igor Geyfman, “jeopardizing the integrity of any environment.” On top of that, startups also face the risk of going over budget. In Building the Foundation for Your Cloud Environment, the second installment of The Startup Guide to AWS Services free video series, Geyfman explains how to meet these challenges.

Smart startup founders will push for innovation without surrendering control of their product, data, or funding. By managing permissions, for example, you can grant access and still protect your data. You also need to establish secure production environments. Geyfman sets up a basic structure, aiming to maintain transparency while boosting productivity. As a first step in account governance, he advises startups to separate their development accounts from their production accounts.

“I’ve worked with many organizations that have had to choose between innovating faster and maintaining control over costs, compliance, and security,” Geyfman says. Instead, he recommends using the relevant services to achieve both.

AWS Organizations automates an account and its environment. You can use this service to set up various AWS accounts, divided between production and nonproduction environments. Then, AWS Single Sign-On centralizes the access to these accounts. That way, a particular user can sign in just once to gain access to all their assigned accounts in one place.

AWS Control Tower can help you build infrastructure quickly, setting up a secure multi-account AWS environment called a landing zone. This process uses AWS Organization and Single Sign-On, Geyfman explains, “with governance and best practices built in.”

Meanwhile, to oversee and control costs, you should track your finances continuously. First, AWS Budgets helps you plan for your cloud costs and track them as they occur. Users can set up budgets, stay informed of costs and usage, and set up alerts in case those expenditures exceed a given threshold—which prevents unexpected monthly bills. Then, AWS Cost Explorer is an analytic tool that allows you to track costs and usage overall, or dig into specific details, to control those expenditures.

AWS can help you maintain transparency and control costs, even while accelerating productivity. That’s the key to building a cloud foundation that allows for growth.

Related Resources

Securing Digital Interactions Within the Pharmaceutical Supply Chain With Spherity

AWS Editorial Team — Thu, 05 May 2022 10:11:15 +0000

Enacted in November 2013, the Drug Supply Chain Security Act (DSCSA) mandates the development of an interoperable electronic system that will enhance the safety and efficiency of the US pharmaceutical supply chain, and ultimately protect patients’ health. Achieving this milestone will require US pharmaceutical trading partners to establish that they only interact with other trading partners who have been properly authorized. This means each trading partner must hold a valid state-issued license or a current registration with the Food and Drug Administration (FDA).

As of early 2022, U.S. pharmaceutical supply chain actors have no established mechanism to validate their counterparty’s trading authorization. With more than 60,000 active trading partners involved in the US Life Sciences industry and stakeholders’ aspiration of responding to data requests in under one minute, the need for an electronic solution that enables regulatory compliance by 2023 is pressing.

Under DSCSA, industry stakeholders expect trading partners to provide a response to a product verification request within 1 minute of receipt of the request.

How pharmaceutical supply chain interactions take place today

As with every supply chain, the pharmaceutical industry has many stakeholders with different needs, including several types and tiers of suppliers, some well-known manufacturers, small and big wholesalers, and a large number of dispensers. There are also several types of interactions and business cases happening concurrently within the pharmaceutical supply chain.

To conceptualize a solution, Spherity chose the “saleable returns” process, which involves the re-sale of legitimate products that have been returned to the seller, e.g. a wholesaler. U.S. pharmaceutical supply chain actors use a routing system provided by so-called Verification Router Services (VRS) to manage acceptance, formatting, and delivery of verification requests and responses for saleable returns based on the GS1 Lightweight Messaging Standard.

A considerable challenge in such business interactions is to be sure about the true identity of previously unknown entities (indirect trading partners) because there is no physical or digital proof of a trading partner at the other end of the digital interaction.

To address today’s challenges in identity authentication of trading partners and the industry’s aspiration of an automated response within 1 minute, Spherity has released CARO, a cloud-based Software as a Service (SaaS solution applying decentralized identity technology to the VRS solutions operating in the U.S. pharmaceutical market. API-facilitated integration of VRS providers allows for little to no changes to existing processes and avoids integration efforts by trading partners.

Spherity’s Credentialing Service CARO – Integration of decentralized identity SaaS with existing business logic leveraging established standards

Integrating decentralized digital identity with existing business processes

The four key components of this solution comprise:

Supply chain actors
Regulatory trading authorization
Existing VRS already in operation
GS1 Lightweight Messaging Standard used by VRS

Spherity’s CARO is based on W3C standard Decentralized Identifiers (DID) to allocate a unique ID to a business entity. By leveraging this decentralized identity technology, a trading partner is able to prove their organizational identity and trading authorization using so-called verifiable credentials (VC). The integrated VRS, that acts on behalf of the trading partner, can use those VC within their message exchange to facilitate product verifications for saleable returns.

Spherity’s Credentialing Service CARO enables trading partners to communicate their own authorized status electronically in every supply chain interaction in an efficient, cryptographically secure, and machine-verifiable way. CARO minimizes the complexity of DSCSA Authorized Trading Partner (ATP) compliance and enables real-time automated verifications in direct and indirect business interactions.

CARO is already integrated with major pioneering VRS providers as well as the forward-thinking regulatory compliance expert Legisym. The latter performs due diligence on trading partners’ licenses and other required documentation in a one-time on-boarding process. Legisym then also acts as a Credential Issuer by transforming the inspected evidence into digital VC that are the foundation of the automated ATP status exchanges between training partners. Any issued VC are stored in CARO’s secure digital wallet. Using this technology enables US pharmaceutical supply chain actors to interact with digital trust.

Spherity’s credentialing solution gives companies a way to prove who they are in the digital sphere using signed, sealed, and verified data with an easy to adopt SaaS solution.

Spherity’s solution follows the Open Credentialing Initiative (OCI) specifications. The Open Credentialing Initiative was co-founded by Spherity and supports the industry in standardizing credential issuance processes, digital wallet conformance and interoperability among service providers.

Decentralized Identities and Verifiable Credentials are managed within the AWS cloud

Forgoing uptime and security is not an option when providing a solution to the US Life Sciences market. With AWS, Spherity is able to leverage a wide variety of technologies and services to provide response times to trading partners in under a second for every product verification request.

CARO allows trading partners to manage the VCs connected to their DIDs for authentication purposes as described. These VCs must be stored in a secure way. AWS supports Spherity with their Multi-Tenant AWS Key Management Service (AWS KMS) to manage keypairs for DIDs and also helps to save VC data securely at REST.

To sufficiently handle all incoming and outgoing requests, scaling is guaranteed through Amazon Elastic Kubernetes Service (Amazon EKS) for the entire Credential Service within their parts to provide trading partners, credential issuers, and VRS providers fast processing.

Long-term storage is also essential for Spherity to fulfill regulatory requirements for conserving credentials and verification transaction records for as long as needed. To this end, Spherity could leverage multiple options to store transaction reports for our customers for multiple years with Amazon Simple Storage Service (Amazon S3) with S3 Glacier as an example.

Spherity Credentialing Service – Uses AWS components to run a secure and performant service for the U.S. pharmaceutical industry

Spherity’s Credentialing Service is designed in compliance with ISO/IEC 27001 and GDPR concepts. The service solution is fully aligned with Open Credentialing Initiative (OCI) specifications. The SaaS solution has proven its high security standard and vulnerability resistance by running through SAP’s cloud solution certification program, becoming SAP-certified, and passing a security test by Veracode with flying colors.

Conclusion: Adopting decentralized identity technology for establishing trust and data security

There are many key takeaways from Spherity’s use case that startups may apply to their own operations.

Fast and easy adoption – It is crucial to approach the implementation of any new process or solution with the overall intention of fast and easy adoption. This means integrating with as many existing business processes as possible, rather than a start-from-scratch approach. This will not only reduce internal lag times around training and education, but it will also minimize opposition from different business stakeholders.
Ecosystem of innovators – Building an ecosystem of stakeholders who are aligned and willing to adopt the technology is necessary. One innovative stakeholder alone attempting to adopt decentralized identity technology creates value in very few use cases. To ensure interoperability of your solution, build on open standards to avoid any vendor lock-ins and work with other interoperable digital identity solution providers to enable the ecosystem to flourish.
Trusted and secure digital interaction – Building on open and global standards (W3C-specified Decentralized Identifiers (DIDs), Verifiable Credentials (VCs), and GS1 Messaging) within AWS Cloud allows companies to empower business relations in a secure and tamper-proof way. Decentralized technology establishes a network effect of trust and data security between business partners within the ecosystem.

Learn more: AWS KMS, Amazon EKS, Amazon S3

Niclas Mietz is a Senior DevOps Engineer with a focus on Distributed Ledger Technologies. He has been building and running different technology stacks for over 10 years. At Spherity, his curiosity encourages customers and engineers to unlock the value of decentralized identities and verifiable credentials on different platforms. If you want to show him something, be sure it has an API and is tested. Follow him on Twitter @solidnerd.

About Spherity
Spherity is a German software provider bringing secure and decentralized identity management solutions to enterprises, machines, products, data and even algorithms. Spherity provides the enabling technology to digitalize and automate compliance processes in highly regulated industries. Spherity’s products empower cyber security, efficiency and data interoperability among digital value chains. Spherity is certified according to the information security standard ISO 27001.

A Startup’s Guide to AWS Services Series 1: Making Security the Cornerstone

AWS Editorial Team — Wed, 04 May 2022 13:05:26 +0000

Building a strong security foundation

Click image to watch video

The security of our customers is the top priority for AWS, and especially true for startups in their initial stages of operation. As an AWS customer, you get the most flexible and secure cloud computing environment available today and benefit from the AWS data centers and network that are architected to protect your information, identities, applications, and devices. Security is a shared responsibility between AWS and the customer, and we want to help ensure you put the right security controls in place for your AWS accounts and workloads as you build and grow.

Building a Strong Security Foundation is the first installment in The Startup Guide to AWS Services free video series. In this seven video series, you’ll learn how AWS services can help you protect both your AWS accounts and your workloads, and stay secure while building quickly. This first video discusses some example baseline controls ranging from access management and secrets management to data and traffic encryption, web application firewalls, and threat detection.

For the full set of controls, we encourage all startups to implement the AWS Startup Security Baseline (AWS SSB). The controls in this guide can be implemented quickly and easily without requiring any additional security knowledge, allowing you to achieve a very strong initial security posture on AWS.

By following the guidance above, security becomes a cornerstone of your development and operation, creating a strong foundation from which you can evolve and scale as you grow your business.

Related resources

AWS Startup Security Baseline (AWS Prescriptive Guidance)
AWS Foundations: Securing Your AWS Cloud (AWS Training)
Getting Started with AWS Security, Identity, and Compliance (AWS Training)
AWS Security Fundamentals (AWS Training)
Security Engineering on AWS (AWS Training)
Best Practices for Security, Identity, & Compliance (AWS Architecture Center)
Protecting Your Data in the AWS Cloud (AWS Training)

ICYMI: AWS Launches $30M Impact Accelerator, Startup Lofts Reopen, Mistakes Founders Make (and How to Avoid Them), and More

AWS Editorial Team — Tue, 03 May 2022 11:05:32 +0000

Announcements

AWS Launches $30M Impact Accelerator for Underrepresented Founders

The newly-announced program will provide up to $225,000 in cash and credits for early stage startups led by Black, women, Latino, and LGBTQIA+ entrepreneurs, as well as training, mentoring, and technical guidance. Plus, hear from three founders about what access to capital and resources means for the next generation.

From the news:

Welcome Back to the San Francisco AWS Startup Loft

The Startups team announced the re-opening of the San Francisco AWS Startup Loft for April 18, 2022. Share the news with your network and with any burgeoning startups looking for support. The SF AWS Startup Loft is open to all startups building on-cloud and is a free benefit for existing AWS Activate members.

Ship Your MVP in Hours with AWS Amplify Studio

AWS Amplify is excited to announce the launch of Amplify Studios, a visual interface to that lets developers go from a Figma design to a feature-rich, full-stack app in hours.

Resources

SERIES: Ten Mistakes Founders Make on AWS, and How to Avoid Them

We want to help you stay focused on your customers and building features for their products, so our Solutions Architects put together a list of the most common mistakes they see founders make on AWS. Each mistake is paired with advice on how you can avoid them to save time and money.

Startup Spotlights

Datagen Creates High-Fidelity Synthetic Data to Address Human-Centric Problems

When Gil Elbaz and Ofir Zuk founded Datagen in 2018, it was with the purpose of re-inventing the broken process of how clients obtain data for computer vision network training. More specifically, they wanted to bring data simulation to every computer vision team in a continuous and scalable way.

How Navina Leverages the Full AWS Toolkit to Make Data Work for Doctors and Patients

Founded in 2018, Navina is leveraging the full AWS toolkit to improve the human-to-human interactions at the heart of healthcare. “[The result is] a better physician experience,” says Anne Amario, Navina VP of Marketing, as well as “better diagnosis and care.” Learn how Navina is driving better patient outcomes and preserving physicians’ revenues.

Using AWS to Build Tools That Will Design Tomorrow’s Green Infrastructure

Web application Optioneer employs multiple AWS services and is updated daily thanks to its continuous deployment pipeline. However, as it grew in complexity, the team at Continuum Industries realized they needed to overhaul their deployment process to deliver more reliable updates to production, faster. That is when the AWS startup team stepped up to provide additional support.

Lokavant Creates a Paradigm Shift in Clinical Trial Intelligence with AWS

Lokavant is a Clinical Trial Intelligence company with the mission to decrease the time and cost of developing drugs by mitigating operational risk. When developing their products and platform they partner to provide the best environment for building and deploying, without bogging down the business with unnecessary costs and effort. Lokavant quickly realized that AWS could help provide the solutions that they urgently needed.

Featured Videos

How AWS Empowered BlockFi to Scale at Their Own Pace

BlockFi, a crypto services platform operating in the fintech space, offers financial products to retail and institutional investors. After rapidly growing from 200 to 1,000 employees during the pandemic, the company knew they needed a solid infrastructure that would allow them to scale quickly and safely.

Dremio Is Transforming How Companies Analyze Their Data on AWS

Dremio, a SQL lakehouse platform, is a service that enables companies to query the data stored in Amazon Simple Storage Service (Amazon S3) to get fast results and power live dashboards directly on their lake.

DayTwo Is Using Data Analytics from the Gut Microbiome to Further Healthcare

Data infrastructure relies on a variety of data analytics tools and machine-learning capabilities, so DayTwo turned to several of the AWS ecosystem services. Specifically, they are utilizing AWS Lake Formation and AWS Deep Learning Containers in order to analyze large outputs. They’re also relying on Amazon SageMaker to manage all of their machine-learning and AI capabilities.

Improving Team Velocity with Software Engineering Analytics using AWS Redshift

AWS Editorial Team — Fri, 29 Apr 2022 13:25:36 +0000

Guest post by Julius Uy, Vice President of Engineering at Hubble, and Jeff Lee, Data Engineer at Hubble

Hubble, a Construction Management Digital Platform based in Singapore, helps construction companies manage the end-to-end lifecycle of a construction project. Over the past few years, Hubble expanded its team size from around eight in 2019 to 16 in 2020 to 35 in 2021. As the team continues to grow, challenges around productivity bottlenecks have also became a growing issue. We noticed that it became harder and harder for us to unblock engineers and keep the wheels moving. In an effort to make sure we’re able to address this, monitoring things such as burnout, knowledge decay, and defects became important.

To dive deeper on the teams’ performances and to support our employee’s well-being, we decided to use analytics. AWS provides a Data Warehousing solution through Amazon Redshift. If we wanted to make sense of our development data, it was essential for us to pipe in this information to a data warehouse, and we felt Amazon Redshift was the ideal solution for us. Amazon Redshift is performant, price-efficient, and very easy to set up. Because Amazon Redshift integrates very well with existing AWS services, such as Amazon Simple Storage Service (Amazon S3), which serves as our Data Lake, we were able to move fast and analyze data quickly compared to other solutions. Along with Amazon Redshift, we also used GitLab APIs to pull the data, Apache Airflow to orchestrate the data ingestion, and Metabase, as our Business Intelligence tool to display the data. In this article, we will share the data we analyzed and what our infrastructure looks like.

Protecting Against Burnout

Research by Gallup has shown that companies with a highly engaged workforce outperform their peers by as much as 147% earnings per share. That’s 1.47 extra headcount for free for every headcount. One of the major causes of engagement drop is overwork. Microsoft Research has found that 20% of overtime per week leads to a 1.6x higher disengagement rate. To address that, we used “commit” activities within our platform’s data to analyze the density work done inside and outside work hours.

The graph above is actual data from the Hubble platform. Over two years, the amount of work done outside work hours increased from 5% to 15% (a 3x increase). Given this information, we then work with HR to protect our staff’s well-being.

Protecting against knowledge decay

“Bus Factor” is the number of engineers that have to be run over by a bus for the project to fail. The higher the bus factor, the lower the risk is for the company. It is therefore essential for tech executives to have visibility on this. At Hubble, we monitor this by the number of commits each person does throughout his tenure for each project. Names of people and projects have been modified in the example metrics below.

Notice that Project 1 is considered low risk because John, Sheryl, and Sandy are all still in the company and have so many commits in the code that the amount of shared context they have is quite high. However, Project 4 is not as good. If Dawson leaves, Pradeep and Vishal will have a hard time filling his shoes. As the VP of Engineering, I need to de-risk the project by delegating Dawson’s responsibilities to Pradeep and Vishal so that over time, Pradeep and Vishal can handle the workload should Dawson one day decide to move on.

Consider Project 5 however where both Enrique and Carl have already left the company. Here, we see that this project is in great danger. Not only is Pragati the only one left, her leaving or falling sick will just cause the project to stop. With this data, I can now expedite hiring for this project and consider moving people from other projects into Project 5.

Protecting against defects

Another issue that often plagues engineering managers is how they can reduce defect density. A study done by Smartbear and Cisco has shown 200-400 lines of code per commit yield around 70-90% of issues found. Anything more than that reduces the capacity of the code reviewer to identify issues. Hence, it is important for engineers to see how well they perform in this metric. We compute this by the total lines of code changed each week divided by the number of commits made in that week. Here’s what a sample data looks like:

As one might observe, Anthony is doing quite well whereas Bella, Dawson, and Janise need some coaching. This will help Engineering Managers raise this in their one on ones to help improve overall team productivity.

Our software engineering analytics stack

GItlab APIs: Accessing your Gitlab data

Users of Gitlab are given the option to extract relevant data that has been created as a result of the use of its application (creating a new project/repository, creating merge requests etc.). Gitlab has a publicly-available link that documents all available REST APIs that can be called via any compatible functions. All it requires is the API URL (and corresponding query parameters) and an access token that can be created within the Gitlab web application.

In our stack, we mainly use Python to execute these Gitlab API calls. The result of these calls usually come in JSON format. Depending on which fields we wish to store as a dedicated column in a data warehouse table (more on how this is achieved in subsequent sections), we then use various objects in the pandas package to convert the inbound semi-structured data into tabular format (as a DataFrame). This multi-step workflow is managed using Apache Airflow, eventually sinking the data into Hubble’s Data Lake.

Amazon S3: Storing our latest Gitlab data with ease and with low-cost

Now that we have established a script to execute the relevant API calls, we would then require a service to persistently store the inbound data. We opted for Amazon S3 as our cloud storage solution to integrate with downstream applications like Amazon Redshift and Metabase.

Our strategy was to turn each Gitlab API dataset response into its own dedicated table. With that, we used Python scripting to transform the original dataset structure in JSON format into a structured file format like CSV. Following this conversion, these CSV files are then stored in their own S3 sub-directories that will be read by our data warehouse application.

Amazon Redshift: Simple, secure, and fast insight on the cloud

Most data visualizations require a data source that a business intelligence tool would be able to connect to. We have selected Amazon Redshift as it fits perfectly with our existing AWS stack, integrating with our underlying data storage in Amazon S3, its distributed nature and its automated table designed to optimize query speeds.

Our Python script programmatically reads every new CSV that has been deposited in Amazon S3 into the corresponding Amazon Redshift table using the common COPY functions. Amazon Redshift’s COPY command reads the contents of an existing file and writes its contents into a Redshift table. All it requires is the fully-defined file URI, Redshift table name, and the corresponding IAM role that has the required permissions to read from the S3 bucket and write into the Redshift instance.

For each of the Amazon S3 sub-directories, we thus transform all the data that has been stored as S3 objects into data warehouse tables which can now be accessed and analyzed with SQL.

Apache Airflow: Orchestrating the flow of data

In order for us to establish an automatic pipeline that pulls, transforms, and stores data without manual intervention, we used Apache Airflow to orchestrate all required steps for analysis. Airflow is open-source and free to use. Using Python coding, data engineers will be able to chain multiple tasks into a collection known as a Direct Acyclic Graph (DAG). On top of that:

Each DAG can be set to run at a given schedule (in Hubble’s use case, we run this one every day).
Tasks can be configured to only run when one or several previous tasks have been successful (task dependencies).
Email notifications can be set up to inform data engineers whenever a pipeline encounters an error.
Secure storage and retrieval of any secrets or connection parameters required for calling the necessary APIs.

Metabase: Free and Open Source Data Visualization

Our final step in our analytics stack is to integrate a data visualization tool to Amazon Redshift. We used Metabase as our tool of choice because it is a free open-source application (using its basic features are free of charge). It also provides data access control by allowing the creation of different user groups and permission settings to manage which groups have access to which data warehouse schemas or tables. Metabase supports more complex analytics by allowing custom SQL queries to generate required result sets, and simple export functions into CSV and Excel file formats are also available.

Overall, we were quite happy with the results. Thanks to Amazon Redshift, we are now able to easily see the performance of our engineering team and be more data-driven in how we approach our execution. With our setup, we are now able to act on problems before they happen. For example, we are able to proactively monitor which projects are at risk, therefore allowing us to hire more engineers to sandbag potential departures. Likewise, because we’re able to nudge the engineers to keep their commit size within the ideal lines of code, we have also seen our Change Failure Rate improve by 300% from July to November. We felt that it is tremendously important to measure what matters, and AWS made this seamless for us to do.

Advice to Other Startups

One might think that a startup is always strapped with resources. Although there are stacks of things to do, it is important for the team to take a step back and evaluate what work will yield the biggest delta in terms of where we are to where we ought to be. It is likewise important for startups to have a close-knit relationship between departments; the engineering team was able to work with HR to understand key HR metrics that were otherwise difficult to analyze manually. There is nevert really a “right time” to start doing Software Engineering Analytics, but that right time may come when it is the highest item in the overall people operations priority list.

How Navina Leverages the Full AWS Toolkit to Make Data Work for Doctors and Patients

AWS Editorial Team — Wed, 27 Apr 2022 12:05:42 +0000

Healthcare is deeply human, and each successful doctor-patient interaction is unique. Some patients feel the most reassured by a straight-forward physician in a white coat; others respond better to a warm and humorous bedside manner. But if there’s one thing nobody wants their doctor to be, it’s overwhelmed.

In addition to the detrimental effects on the doctors themselves, rampant burnout among healthcare providers puts patients at risk. “It’s almost inevitable that doctors will miss important things,” says Anne Amario, VP of Marketing at Navina. “That includes inevitably missing data that’s hiding in those overstuffed patient records, and missing significant diagnoses, too. That’s where Navina comes in; only AI can make this data work for physicians and not against them.” Founded in 2018, Navina is leveraging the full AWS toolkit to improve the human-to-human interactions at the heart of healthcare. “[The result is] a better physician experience,” says Amario, as well as “better diagnosis and care.”

A 2018 survey found 78 percent of doctors reported feeling burned out. If the workings of the human body weren’t complicated enough, healthcare providers today must navigate a complex terrain of patient data and clinical research while carrying ever-growing patient loads. The same survey found that doctors saw more than 20 patients each day on average. As the pressures of the COVID-19 pandemic have driven healthcare providers across all levels of care out of the industry, the trend toward burnout has only gotten worse.

In an effort to cut administrative costs and make important patient data more accessible to distributed networks of providers, the passage of the Affordable Care Act in 2009 mandated the digitization of patient records. “But having all the information doesn’t mean the information is readily available for physicians to use,” Amario says. Patient records remain complex, unstructured, and unprocessed. Parsing them all is more than most overwhelmed doctors can manage in the scant minutes before each patient meeting.

Navina changes that by providing physicians with the tools to control their data in every interaction with their patient. Using proprietary transformers-based Natural Language Processing (NLP) models, Navina’s AI can analyze and filter this unprecedented volume of available information. Navina integrates the data from electronic medical records (EMRs) and other sources, surfaces what’s important, identifies potential quality and risk gaps requiring attention – and empowers physicians to take action and to document on the spot. It uses the Amazon Simple Storage Service (Amazon S3) to manage millions of patient records. Meanwhile, Amazon Textract and Amazon Comprehend scan clinical documents including hospital and consultation notes, cutting through medical jargon to identify issues that require physicians’ attention, like missing labs, abnormal results, and potentially missed diagnoses.

And, as Amario explains, “Navina is shifting the patient-doctor interaction from reactive to proactive, from “What’s wrong with you today?’ to ‘How can can we optimize your health?’” so users feel safer when they have Navina’s support. “It’s like having another set of eyes, or another virtual doctor that has all the time in the world to read through every single document and alert them to what could have clinical significance.” And in a value-based care environment, where physicians are financially reimbursed for their performance and missed diagnoses mean lost revenue as well as poor patient outcomes, Navina helps providers avoid millions of dollars in losses. One physician group that has integrated Navina into their practice has pulled in $45,000 in additional revenue per provider per year, adding up to more than $1 million in revenue they would have otherwise foregone.

Navina’s user base is growing fast, having started off with having started off with hundreds of physicians and now serving over 3,000 users in dozens of clinics throughout the United States eager to put the tool to work. The growth of the value-based care model throughout the healthcare industry means the most exciting developments are likely still to come. The machine learning (ML) infrastructure of AWS allows Navina to accelerate the training process for its proprietary ML and NLP models, so the company can scale quickly and streamline experiments as needed.

Looking towards the future, Amario says, “The irony is that the entire ecosystem is focused on generating more and more data about patients, but in this way everyone is adding to the workload and nobody is focused on how the data will be digested. Navina is staying ahead of the curve, with an eye toward how physicians can make the most of all this data, and creating space for more meaningful patient interactions.”

Automating Unstructured Data Processing with Amazon SageMaker

AWS Editorial Team — Tue, 26 Apr 2022 13:06:59 +0000

by Nikhil Dinesh, Head of Startup Business Development, DACH Region, AWS, and Sayon Saha, Machine Learning Specialist Solutions Architect, AWS

Unstructured data such as images, video, and text that appear on E-commerce product listings have a significant impact on conversion rate. A study by eBay research found that super-sized images can increase conversion by 15.3%, with other factors such as photo count and item condition playing a significant role. Marketplaces and sellers must optimize conversion based on an open-ended set of factors determined by marketing teams. The use of data science and machine learning (ML) to address this problem is not new: AWS has created several services to help with the undifferentiated aspects of ML, such as Amazon Rekognition (for images and video), Amazon Comprehend (for text), Amazon SageMaker (for model development and deployment), and Amazon SageMaker GroundTruth (for data annotation).

Super.AI, a Berlin-based startup, believes that there is a significant opportunity to assemble these building blocks in the right way and with the right user experience toward what they call Unstructured Data Processing (UDP) in various industries. According to Gartner, 80% of data in a typical enterprise is unstructured. Super.AI’s platform extracts actionable information from unstructured data, allowing enterprises to automate complex business processes. According to Brad Cordova, serial AI entrepreneur and Founder/CEO of super.AI: “Customers across E-commerce, TIC (Testing, Inspection, and Certification) Services, Insurance, Healthcare, Manufacturing, and Agriculture are using super.AI platform to automate complex uses cases such a product listing quality assessment, visual inspection, vehicle damage detection, and crop yield assessment. Our customers are achieving significant ROI via reduced time and costs, fewer errors, and improved customer satisfaction.”

This piece will show you where data ingestion, pre-labeling, the active learning pipeline, and real-time assisted labeling falls in the super.AI architecture on AWS, followed by a discussion of the goals, risks, and where we see opportunities for improvement.

Product image tagging in an eCommerce setting

The detection of serial and model numbers, with options for manual correction.

Super.AI’s Unstructured Data Processing Platform

The super.AI platform helps customers to transform processes involving unstructured data such as images, videos, text, documents, and audio and automate them using a combination of AI, software, and humans. This super.AI Product Image Categorization Demo shows how super.AI product image categorization can help retailers increase website conversations.

Active Learning and Pre-Labels

Super.AI customers requested a more efficient, highly accurate labeling mechanism. So, they recently released a new feature called Active Learning and Pre-labeling, where the pipeline pre-processes data points using an ML model running on SageMaker. This solution prioritizes labeling data points that are most useful to the model. The ML model is run over all the uploaded data points to generate an output, e.g., a confidence score, used to serve the data points in a prioritized way. Pre-labels are generated where possible and served to the human labeler for review or editing.

The pipeline then scales with demand. Customers can upload data via API (or UI) and apply multiple models for active learning and pre-labeling. Customers can choose from a selection of models provided by super.AI or bring their own model. Super.AI uses the confidence score generated by the ML model to prioritize the data points and serve them more efficiently. Where necessary, human labelers can use pre-labels generated by the system to label the data by hand accurately.

Real-time Assisted Labeling

The platform leverages a serverless architecture with SageMaker. The customers must have this service in real-time for labeling their images. The tool leverages AWS Lambda in combination with Amazon SageMaker Endpoints to serve concurrent requests in real-time with response time under 10 seconds. You can explore the Image tagging application from super.AI in their online documentation.

Building an Active Learning Pipeline with Amazon SageMaker GroundTruth

SageMaker Ground Truth is a managed data labeling service to build large-scale accurately labeled ML datasets with several workforce options. Along with the various built-in and custom data-labeling workflows for text, image, videos, and 3D point clouds, it allows you to build an automated data labeling pipeline with active learning by automatically annotating objects with relevant ML models and assigning objects with lower confidence for human annotation.

The first step of the pipeline includes SageMaker Ground Truth sending a random sample of datasets for human annotation to train and validate the model used for auto-labeling. The trained model’s outputs confidence score and quality metric in the validation data are compared against the threshold for deciding on the quality labels for annotating the rest of the dataset. Depending on whether the confidence score meets the desired threshold, either an object is considered auto-labeled or sent to the human workforce for annotation. These annotations, in turn, are used to update and improve the auto-labeling model. This active learning pipeline continues to process until the required dataset is fully labeled or another stopping condition is met (Learn more in this article to Automate Data Labeling). The Active Learning process is illustrated in the following diagram:

While many AWS customers simply use its built-in ML models, SageMaker Ground Truth allows you to bring your own if you have a custom use case. You can read more about it in the blog “Bring Your Own Model for Amazon SageMaker Labeling Workflows with Active Learning.”

The active learning technique makes the data labeling process much faster by identifying the subset of data that your labelers should label. It also reduces the workforce cost significantly by keeping the accuracy of annotations high. You can read about an example use-case of object detection job with automated data labeling in the blog “Annotate Data for Less with Amazon SageMaker Ground Truth and Automated Data Labeling.”

Conclusion

For the last few years, robotic process automation (RPA) has been one of the fastest-growing software categories as enterprises strive to achieve digital transformation. However, 80% of the enterprise data is unstructured and off-limits to automation. Emerging Unstructured Data Processing solutions from companies like super.AI that leverage AWS ML services are helping enterprises greatly expand the scope of automation by extracting actionable information from unstructured data – images, videos, audio, documents, and text. Such platforms can address a wide range of uses case from visual inspection to the online product listing quality assessment with minimal human intervention. Early adapters of such platforms are gaining a competitive advantage which drive down costs, reduce errors, and deliver a differentiated customer experience.

Ship Your MVP in Hours with AWS Amplify Studio

AWS Editorial Team — Fri, 22 Apr 2022 12:23:12 +0000

As a startup founder, you’re creating something from an idea. The need to push an MVP to market quickly looms large over the early stages of the startup lifecycle and may be the most pivotal factor in your organization’s success. When you’re ready to build—and no matter how inspiring your idea—you have to balance choices between budget, resources, and time: every decision you make during the development process can impact long-term productivity and product functionality; likewise, any compromise in the quality of the user experience puts your success at risk.

AWS accelerates front-end web and mobile development with service like AWS Amplify, a set of purpose-built tools and features that let developers quickly and easily build full-stack applications on AWS so you can get to market faster. As your business grows, Amplify can be extended with other AWS services and tools to support custom features and app requirements. Amplify is built on top of AWS serverless services, so you benefit from the reliability of AWS infrastructure to deliver secure, scalable apps from day one. And today, AWS Amplify is excited to announce the launch of Amplify Studios, a visual interface to that lets developers go from a Figma design to a feature-rich, full-stack app in hours.

Simplify front-end web and mobile development

To remain focused on creating a differentiating end-user experience, front-end web and mobile developers seek ways to be more productive building their application. Flexibility and control over their app code enables them to envision an app that scales as the business grows, and they prioritize the ability to iterate to meet evolving customer needs, while also integrating with their favorite tools and services. With an MVP, the real goal is getting it out the door as quickly as possible and making sure it’s ready to scale.

AWS Amplify launched Amplify Studio to help startups get to market quickly and leverage the breadth and depth of AWS services as their business needs and requirements evolve. With Amplify Studio, you can quickly create your app frontend UI, configure your backend capabilities, and connect them together with minimal coding. AWS Amplify Studio builds on Amplify’s capabilities, enabling developers to easily configure their backend data model, authentication, authorization, and more. All backend resources created by Studio are usable with the Amplify CLI. Once your application is launched, developers can invite application operators to access the user and content management view to manage users and app content without relying on developers to push updates. Let’s take a look under the hood.

Accelerate your front-end UI development

Studio comes with dozens of popular React components—such as buttons, forms, and marketing templates, so you can select and customize them to fit your style guide.

Alternatively, for developers who are working with UI/UX designers, they can import UX designs from the popular design prototyping tool Figma as clean React code for seamless collaboration. Amplify Studio imports all UI and infrastructure artifacts as code, so developers can maintain full control over your app design and behavior.

Amplify Studio facilitates the designer and developer workflow, providing more seamless collaboration by meeting both designers and developers where they like to work. In application design and development, optimizing the designer-developer workflow can be challenging and time-consuming, even though both UX designers and developers are dedicated to delivering exceptional end user experiences. Amplify Studio helps designers and developers collaborate better together, reduce inefficiencies, and create exceptional user experiences by meeting UX designers and developers where they are, centralizing the designer-developer workflow, and providing flexibility and customization.

Host your app

If you’re building a web app or static website, you can use Amplify Hosting to deploy and host any React, Vue, or Next.js web app with built-in continuous integration and continuous deployment (CI/CD) workflows, testing, pull request previews, and custom domains.

Grow on AWS

Amplify is scalable and extensible, which means that you can easily get started using its core use cases, and you won’t hit a wall as your needs grow. Amplify lets you leverage the breadth and depth of 175+ AWS services to add further capabilities to your app. And, you get the reliability, security, and performance at scale that AWS provides.

Here are just a few examples of how other startups are building and growing on AWS using AWS Amplify:

Busby grew its user base by 860 percent without a single glitch: Busby wanted to build an app to help stranded, injured cyclists. Due to a limited budget, the company decided to develop the MVP on its own. To make it happen, Busby needed an end-to-end, easy-to-use, serverless app-building solution. With AWS Amplify, they were able to complete their MVP in three months, saving an estimated six months of development time and over GBP 100,000 in outside development costs.
Credit Genie pushed its consumer-facing app live in under seven days: The FinTech startup leverages data analytics to help borrowers escape the personal debt trap by optimizing their debt to manageable levels. They built a five-page consumer-facing app for their MVP using AWS Amplify in under a week, streamlining their codebase and reducing overhead costs.
All of Us Financial gets unmatched security and scalability with AWS: Due to the sensitive nature of financial trading, All of Us Financial required a solution with maximum security for all transactions performed via their platform. Rather than spending months creating their own backend infrastructure, they utilized the efficient, low-cost, proven-cloud-native solutions of AWS for unmatched security and operations. Building with AWS Amplify enabled their front-end web and mobile developers to easily leverage the power of services, such as Amazon Cognito, to create a secure, scalable, full-stack application.

Build and deploy a React, Flutter, iOS, or Android app on AWS—both a backend with user authentication, a database, and storage and a front-end UI— in hours with minimal coding. Start building your MVP today.

AWS Launches $30 Million Impact Accelerator for Underrepresented Founders

AWS Editorial Team — Wed, 20 Apr 2022 12:32:17 +0000

New program provides $125,000 cash and up to $100,000 in AWS credits for early stage startups led by Black, women, Latino, and LGBTQIA+ founders, as well as training, mentoring, and technical guidance.

In 2018, Claudius Mbemba (mem’buh), a Black man who grew up in Cameroon before moving to the United States, co-founded his first venture scale business—a managed marketplace app for vacation rental cleaning services called Neu. It gave him firsthand experience of the challenges of navigating the investing landscape. As someone from an underrepresented community, he’s seen how the pool of businesses that receive funding is often homogeneous.

“It’s all about the money flow. As soon as investors and those with access to opportunity start investing in a more diverse range of recipients, then we will start moving in the right direction,” said Mbemba. “It’s a self-fulfilling prophecy. You get access to the capital, which creates incentives for investors to open up their networks to you, which can help your business meet the next milestone, which creates success. When people don’t invest in founders from diverse backgrounds, they’re not creating that momentum and not helping fulfill that prophecy. It comes down to that.”

This funding gap, as well as difficulties accessing training, networking, and mentoring opportunities, are just some of the issues faced by underrepresented founders like Mbemba that Amazon Web Services (AWS) is committed to addressing. It’s why AWS is launching a new AWS Impact Accelerator that commits more than $30 million over the next three years to early stage startups led by Black, women, Latino, and LGBTQIA+ founders.

Each qualifying startup receives up to $225,000 in combined cash and credits, training, mentoring, and technical guidance, as well as introductions to Amazon leaders and teams, networking opportunities with investors, and advisory support. Eligible startups can now apply to the first of these program, the AWS Impact Accelerator for Black Founders.

“I’ve been in situations and cohorts where people tell me it’s so easy to raise money,” said Mbemba. “And I’m thinking, ‘That’s not the experience I’m having.’ Friends and family investing doesn’t work for me. I don’t have people who can just invest $100,000 to get my business off the ground.”

To put this into perspective, studies consistently show that Black, women, Latino, and LGBTQIA+ startup founders are underrepresented across the total number of founded and funded startups. In 2021, U.S. startups founded solely by women raised nearly $6.4 billion in venture funding, according to PitchBook data. But that number only accounts for 2% of overall investment dollars that year, which was $330 billion. For example, according to StartOut, an organization that supports LGBTQ+ entrepreneurs, if everyone had equal access to funding and resources, there would be 10 times more LGBTQ+ entrepreneurs in the U.S.

We spoke to Mbemba, an AWS customer, and two other startup founders who work with AWS, about their experiences of building successful businesses as a part of an underrepresented demographic, how strong partnerships and accelerator programs changed the game for them, and why more initiatives like the one from AWS are vital.

Marcus Bullock, CEO, Flikshop

In 2012, Marcus Bullock launched Flikshop, a platform that sends photos delivered as postcards to any person in any prison instantly, with an often difficult topic on his resume—a felony conviction. At the age of 15, Bullock entered a maximum-security prison to begin an eight-year sentence for carjacking a man in a parking lot. During this time, he said he cycled through “anger, frustration, and complete disillusionment.” What kept him afloat was regular correspondence from his mother.

“She wanted me to understand that there was still life to be lived,” said Bullock. “And so, she made a commitment to send me photos and write me letters every day until I came home.”

That constant reminder of life on the other side continued to drive Bullock when he got out. He founded a painting business and used the profits to fund the development of Flikshop, an app built on AWS that lets anyone take a picture, write a message, and send it as a physical postcard to an incarcerated friend or loved one.

Yet Bullock had no technical skills when he left prison. Inmates weren’t allowed to use computers, and when he began his sentence, the internet was still on the perimeter of mainstream culture. His only experience of it was through movies. “You think you understand what it does. But it didn’t really exist when I went it, and when I came home, there it was,” he said. “To go from nothing to suddenly being able to type anything into a search bar, and to have all this information pop up, was just the most mind-blowing experience ever. The whole world had changed.”

His reaction? To dive right in. The more Bullock learned about developing applications for smartphones
and the web, the more he learned what he needed to learn. Questions that tech insiders consider common parlance were alien to him at that point. “What does UX/UI mean? What are wireframes? How do we think about pricing? What is a server? What’s the difference between hosting servers and cloud computing? I had a massive knowledge gap in trying to figure out how to launch,” he said.

It’s a gap that Bullock has been incredibly successful in bridging. Flikshop has now shipped over 700,000 postcards to inmates in jails, prisons, juvenile facilities, and ICE detention centers around the country, and continues to grow at 20% month over month. The company has also developed enterprise partnerships with Fortune 100 companies such as Boeing, Bank of America, and AWS.

Looking back, when Bullock considers what advice he would have given himself when starting the business, one thing stands out—strengthen your network.

“I didn’t realize that having a felony would matter after I had proven I could generate revenue,” he said. “Or that not having access to social capital would differentiate me.”

That’s where accelerator programs and competitions changed the game. In 2018, Flikshop entered two incubator programs and finally developed the traction it needed to grow. Having singer John Legend’s campaign FreeAmerica partner with Bank of America and New Profit to invest in the business also helped.

“You can’t outwork relationships,” said Bullock. “No matter how incredible your product is, no matter how amazing the technology or the tool you’re able to introduce, no matter how incredible your marketing or strategy across platforms—relationships are what matter. It’s the people who become your advocates, supporters, and sponsors. It’s the people who kick down a door for you and say to someone, ‘Hey, you need to figure out a way to collaborate with Marcus.'”

Adriana Vasquez, CEO, Lilu

When Adriana Vazquez started her smart garment company Lilu in 2017, she quickly found that the resources readily available to her Ivy League classmates were far harder for her—a woman and a Latina immigrant—to access. Recent Crunchbase data, for example, shows funding for early stage startups led by Latinx founders has almost completely stalled in the past few years. “Such a small percentage of venture capital funding goes to Black and Latina founders,” said Vazquez. “The odds are really stacked against us.”

But Vazquez persisted. Leaning on technical degrees from both MIT and the University of Pennsylvania, she developed the first Lilu smart bra for breastfeeding mothers. Vazquez had seen so many of her women colleagues deal with the stress, even shame, of pumping—a critical task for any mother who needs to keep up her milk supply—while at work or on the go. Not only that, but most pumps were cumbersome and uncomfortable to use.

“Technology should make people’s lives better, easier, and bring joy into them,” she said. “I was surprised by how technology wasn’t supporting moms as well as I knew it could.”

But Vazquez still needed to prototype and get her product into customers’ hands, all of which required a

great deal of testing and iteration, not to mention time and capital. In the beginning, Lilu was aided by grants and prizes in pitch competitions, but it was the boost it received from larger accelerators, and
funding from America’s Seed Fund, a National Science Foundation program, which Vazquez said really put her company in gear.

“A good accelerator with a good program, really early on, can make such a big difference,” she said. “So many founders don’t have necessary resources, connections, or access to capital. But these are the things that make the impossible possible. We have people innovating from all cultures and parts of the world, and they typically don’t have access.”

Vazquez noted too how AWS, even before the launch of the Impact Accelerator, kept the door open—helping her hone her pitch and connecting her to resources that could move Lilu closer to its goals.

“AWS has always given me a platform to talk about what I’m doing,” she said.

Claudius Mbemba, CTO and co-founder of Spritz (formerly neu)

Claudius Mbemba, who grew up in Cameroon and immigrated to the U.S., saw from a young age how technology could improve lives in his country of birth. He showed an early talent for shaping technology as he saw fit, writing computer programming scripts and even hacking a gaming system his father bought.

“I was fascinated with technology and how it could enable me to access and create the things that I wanted to see in the world,” said Mbemba. That drive eventually led him to major in computer science at Ohio State University, before working as a software engineer. He later cofounded Neu, a business that helped connect vacation rental owners with housecleaners, the way Uber connects drivers and passengers.

“I’ve always had a knack for spotting things that could be better and solving that problem,” he said.

Today, Mbemba is engaged in a huge pivot, transforming Neu into a new business called Spritz—a software as a service (SaaS) platform that provides housecleaning professionals with the tools they need to grow and manage their operations, such as back-office functions, operations, logistics, and administrative tasks.

To do this, the company needed advisors and a broader field of investors who share Mbemba’s experience and deeply understand the business “It’s about having representation and investors and mentors understanding the technology and business model.”

Mbemba points out that immigrants and entrepreneurs share an important quality—making big, bold leaps. He has an adage when it comes to taking the plunge, as he’s done many times in his life.

“The sooner you start, the sooner you finish,” he said. “Often, prospective entrepreneurs and business leaders will delay because they don’t know how exactly to make a change. On the startup journey, you will encounter ‘unknown unknowns.’ Those unknowns are best solved by just doing the thing.”

AWS Impact Accelerator

The AWS Impact Accelerator gives high-potential, pre-seed startups the tools and knowledge to reach key milestones such as raising funds or being accepted to a seed-stage accelerator program, while creating powerful solutions in the cloud. Participants accepted onto the eight-week program will create a wide-ranging, personalized training curriculum from dozens of available sessions delivered by AWS startup experts and guest speakers, and will also learn how to use Amazon processes such as “two-way door decision making” and “working backwards” to drive day-to-day decisions and build nimble, innovative teams.

Applications open today for the first AWS Impact Accelerator for Black Founders, with the program kicking off in June for U.S. based startups. The first AWS Impact Accelerator for Women Founders will take place in the second half of the year for U.S.-based startups, and the AWS Impact Accelerators for LGBTQIA+ Founders and Latino Founders respectively will follow in 2023.

Read the press release for more information.

Video: DayTwo Is Using Data Analytics from the Gut Microbiome to Further Healthcare

AWS Editorial Team — Tue, 19 Apr 2022 14:20:34 +0000

DayTwo is a microbiome discovery platform working to find new links between the bacteria living in our bodies and different health conditions. “We wanted to take our analysis one step further and to dive into a higher resolution of the gut microbiome analytics,” explains Adi Lev, DayTwo’s Vice President of Research and Development. “This meant we needed to move from the bacteria-species level to the gene level.”

Because data infrastructure relies on a variety of data analytics tools and machine-learning capabilities, DayTwo turned to several of the AWS ecosystem services. Specifically, they are utilizing AWS Lake Formation and AWS Deep Learning Containers in order to analyze large outputs. They’re also relying on Amazon SageMaker to manage all of their machine-learning and AI capabilities.

“The AWS infrastructure offers different services—from pipeline orchestration to distributive computing to building machine-learning models—that can help us with building this pipeline end-to-end,” says Lev. “By being able to collate between features from our gut microbiome to different health conditions, we can actually find new links, new biomarkers, and make them the seed of what can become new drugs for drug discovery, early diagnostics, and other healthcare solutions.”

To learn more about DayTwo’s missions, check out their speech for re:Invent 2021.

Spotlight: Datagen Creates High-Fidelity Synthetic Data to Address Human-Centric Problems

AWS Editorial Team — Thu, 14 Apr 2022 10:31:45 +0000

Because AI model performance relies on both the quality of the model and the quality of the data used to train it, it’s essential to have a large quantity of good data—and it’s often challenging to collect as much as needed. Real-world data also tends to be problematic when it comes to speed of acquisition, precision, expense, and bias. “Someone will collect [real-world] data of different identities—for example, for faces—and they will not collect enough of certain ethnicities , or ages, or gender,” explains Shay Navon, Datagen’s Senior Product Marketing Manager. “And then you get this bias.”

In order to help computer vision teams fight bias, Datagen offers a unique way to generate data using computer algorithms. Its synthetic data is akin to real-world data both statistically and mathematically, but can be generated quickly, with less expense, and is exempt from human error. Instead of tasking a human with the chore of gathering and annotating data manually – a labor intensive task—which requires them to take a photo of a face and then label its features by hand—the synthetic data is generated on a massive scale, with built-in ground truth annotations , such as eye direction, that would be impossible for a human to determine. The result is more accurate and detailed data annotation without the challenge of manual tagging.

“We are simulating the world to bring AI to production faster,” says Karine Regev, Datagen’s VP of Marketing. “Bringing AI to production is by itself an unsolved challenge for most companies out there, so we are making it more professional, more accurate, solving problems like privacy, solving problems like bias in the data which are the largest bottlenecks in modern AI.”

Datagen offers clients a self-service platform that uses 3D simulations to train their algorithms. “In order to train a model, you need millions of different images,” says Regev. “And this is exactly where we fit in. [Datagen customers] have the ability to control the scenes, the ability to control the background, the different modalities, the different labels that you need, the lighting, the gender, ethnicity, everything.”

Datagen Platform, self-service synthetic data generation

In addition to generating diverse data that looks real, scales, and is pixel-perfect, Datagen offers its customers complete confidentiality. “It’s fully privacy compliant, as the data contains zero PII (Personally Identifiable Information),” says Shay Navon about the synthetic data. “Nobody can say, ‘This is someone that we are using that is a problem privacy-wise.’ Our human-centric expertise and data focuses on several domains, from facial landmarks detection, gaze estimation, and expression analysis to full human body poses, body parts like eyes, hands, etc.”

In the very near future, it’s predicted that it will be more common to train models with synthetic data than to collect it from real-world sources. In accordance, Datagen has been growing rapidly, expanding from around 40 employees to nearly 100 over the last nine months. “We are working with some of the largest tech companies in the world in different verticals,” says Regev. “Solving different use cases from AR/VR/Metaverse to driver monitoring for in-cabin automotive, to home security and smart offices.”

Datagen Platform: Home Security

In order to meet this new demand, Datagen decided to switch to cloud architecture. Their priority was to scale using the latest GPU models. After an in-depth analysis of cloud providers, they turned to AWS, determined to develop their system on top of Kubernetes. Datagen designed a custom scheduling software system called Agni that integrates with Elastic Kubernetes Service (Amazon EKS) and uses Kubernetes auto scaling and AWS Auto Scaling Groups.

Agni—and the entire Datagen data-generation platform—now relies on CPU and GPU spot instances, which has helped them to reduce costs and build a more efficient system. It also enables them to maintain a relatively small system that can dynamically grow to hundreds of thousands of concurrent jobs and shrink on demand, resulting in a self-service platform hosted by AWS.

Looking forward, the Datagen team predicts that the need for synthetic data will continue to grow. “We’re seeing a lot of demand, both in the traction and prospects, the need for thought leadership, the need for technology, and a solution like ours that can actually lead the conversation when it comes to synthetic data,” says Regev.

Video: Dremio Is Transforming How Companies Analyze Their Data on AWS

AWS Editorial Team — Wed, 13 Apr 2022 10:30:18 +0000

“A lot of companies have data on AWS and need ways to analyze it,” says Tomer Shiran, Dremio’s Founder and CPO. “And moving the data is not a very compelling idea.”

In order to remedy this, Dremio built a service on AWS that makes it easy for clients to connect to their data and query it directly. “Companies don’t have to install software,” says Shiran. “They don’t have to upgrade anything or monitor anything. They just log into app.dremio.cloud, which is an AWS service, and start using it.”

Dremio also offers mission-critical BI dashboards on their data lakes. This drives agility, enables self-service, and makes it easier to work with data and democratize data within organizations.

“Our customers love being able to use Dremio with the AWS Glue catalog and with lake formation and with the data that they have in S3,” says Shiran, adding that Dremio works with many different Amazon product teams and integrates a variety of different services. “That really tight integration and collaboration that we have provides a lot of value for our customers.”

Using AWS to Build Tools That Will Design Tomorrow’s Green Infrastructure

AWS Editorial Team — Tue, 12 Apr 2022 14:25:50 +0000

Guest post by Connor Philip, DevOps Engineer Continuum Industries, with Amir Majlesi, Principal Manager at AWS EMEA Prototyping Labs and Bigad Soleiman, Lead Prototyping Architect

Continuum Industries is an Edinburgh-based startup developing AI tools that enable engineers to rapidly create and explore design options for infrastructure projects. We’re on a mission to help accelerate the path to Net Zero by speeding up the early-stage design of large, complex infrastructure projects. To date our web application, Optioneer, has successfully helped support the design of major water and offshore wind transmission projects. The projects we work on are built-in and influence the real world; they help secure our fresh-water supply and help connect offshore wind farms to the mainland.

Optioneer employs multiple AWS services and is updated daily thanks to our continuous deployment pipeline. However, as it grew in complexity, we realized we needed to overhaul our deployment process to allow us to deliver more reliable updates to production, faster. We had been aware of some issues but were reluctant to start making changes. Our concern was any change would become a big resource drain on our devs and drag on after multiple deadlines.

That is when the AWS startup team provided additional support. When we explained our challenges, and they suggested working together with AWS Prototyping Labs to build a system to handle canary deployments with our current tech stack. A prototyping engagement with AWS is focused on co-development of an end-to-end solution together with AWS experts whereby we can rapidly design, build and test in an agile environment following AWS best practices. We agreed on a timeline of 5 weeks. While we weren’t sure how much we could achieve in such a short span of time, we were excited about the prospect of working with AWS experts.

Developing the prototype

Our team of three, a DevOps and AI engineer from Continuum Industries and a prototype architect from AWS started implementing the canary deployment system by slotting it into our current deployment cadence whilst smoothly fitting in the validation of complex configurations only available on our production system. It was important to us to take full advantage of the prototype architect support to learn best practices when using AWS to build this system.

Our basic deployment diagram before prototyping

Daily stand-ups between the CI and AWS team members and weekly reviews of progress helped us make adjustments in real time and share learnings that other team members could utilize in their own tasks.

Once we determined the core components which provided more confidence in our vision, progress quickly picked up. The individual pieces of the puzzle started to come together:

Our DevOps team member created the deployment workflow.
Our AWS prototyping architect created the core business logic.
Our AI engineer put forward the overall Step Functions design.

A prototyping approach helped us push forward quickly, with the AWS prototyping architect guiding us through design choices and demonstrating how various AWS services would work together to meet our unique requirements.

Deployment diagram of what we created after the prototyping period

Our AWS solutions architect suggested we use AWS Step Functions, a low-code, visual workflow service, which became the core of our new deployment system. Its built-in ability to orchestrate AWS Lambda functions, Amazon Elastic Kubernetes Service (Amazon EKS) job executions and various AWS services enabled a custom, flexbile integration with our system at every level. Its modular nature makes it easily extendable with the potential of being service-tailored as needed. During development we were also introduced to AWS Step Function Workflow Studio, which was a vital visual tool in the process of implementing workflows as well as including them in our Infrastructure as a Code (IaC) framework, Terraform. The workflow we designed deploys the latest version of our AI service to our production cluster, where we shift traffic over to the new version at a more efficient pace. Throughout this process, we also perform a set of custom checks against on our services, leveraging AWS Cloudwatch, to ensure it’s safe to proceed. Once 100% of the traffic is routed to the new service, we remove the current stable version and replace it with the updated stable one.

When the new version deployed successfully, we initiated the “config validation process.” This automatically generated a schema for the configuration our AI service expects, and verified that configurations already present in our production database weren’t in violation. Were that the case, a Cloudwatch Alarm would fire off, the rollback workflow would kick in, and a detailed report with useful links to logs and a dashboard would be sent to our internal dedicated Slack channel to ensure maximum visibility and swift debugging. The integration with Slack was made possible thanks to an extendable notification module built on top of Amazon Simple Notification Service (Amazon SNS), Amazon Simple Queue Service (Amazon SQS) and Lambda.

These workflows enable our engineers to perform safer deployments and migrations without impacting the production stack and act immediately with relevant data at hand when needed.

Conclusion

Together with the AWS prototyping team, we successfully co-developed a working prototype within the five-week goal. This partnership provided a scalable architecture our team could build upon, allowing us to iterate faster and have greater control for running different versions of the algorithms for different users. The serverless architecture is based on Step Functions, aiming to orchestrate several small, reusable logics in order to provide full flexibility to meet bespoke deployment requirements but also allowing future extensions.

The AWS Prototyping Team didn’t just help us build the architecture, they transferred their knowledge in a way that empowered us to continue developing on our own. At the end of the five weeks, we achieved our goals:

Our new deployment system is capable of detecting issues within the service and production configurations during deployment and roll back when required.
We shared the knowledge gained and learned a lot from the insights AWS prototyping architect gave us.
We went further and managed to add a robust monitoring and alarm system during the prototyping period.

This success wouldn’t have been possible without the team having a prototyping mindset from the outset and without the great help from our AWS prototyping architect. Adopting a prototyping mindset encouraged us to:

Communicate often and ask lots of questions,
Adapt our original plan when circumstances change,
Have an agile mindset to rapidly design, build, test and learn from failures,
Always think about the end goal when prioritizing tasks,
And step out of our comfort zones.

If you find yourself stuck in a situation where you know a change is required, but you’re not quite sure of the correct solution, I recommend getting the AWS prototyping team together to focus on the task.

Connor Philip: Connor is a DevOps Engineer at Continuum Industries. He develops the systems that allow the entire technical team to create tools which will design tomorrow’s green linear infrastructure.

Amir Majlesi: Amir is a Principal Manager at AWS EMEA Prototyping Labs. He supports customers with exploration, ideation, engineering and development of state-of-the-art solutions using emerging technologies such as IoT, Analytics, AI/ML & Serverless.

Bigad Soleiman: Bigad is a Lead Prototyping Architect with an extensive software engineering background at Amazon Web Services. Leading customers through tough business challenges involving Serverless, DevOps, and other cutting edge technologies.

Lokavant Creates a Paradigm Shift in Clinical Trial Intelligence with AWS

AWS Editorial Team — Thu, 07 Apr 2022 14:39:45 +0000

Guest post by Brian Monroe, Lokavant Senior Director IT and DevOps, and Rohit Nambisan, Lokavant CEO.

One in six clinical trials of investigational drugs fail regulatory review by agencies such as the Federal Drug Administration (FDA) and the European Medicines Agency (EMA) due to issues stemming from the way data are managed and interpreted — problems that have nothing to do with the safety or efficacy of a drug candidate. Drugs that fail in regulatory review are marred by delayed timelines and budget overruns; indeed, the average cost to develop an approved drug has more than doubled to $2.6 billion over the last ten years. The consequences of these operational failures can cost tens of millions of dollars for sponsors of clinical trials and their Contract Research Organizations (CROs) in direct cost and much more in opportunity cost from lost sales due to late market entry. Those who suffer most, however, are the patients, who have access to fewer treatment options and are burdened by higher drug prices.

Lokavant is a Clinical Trial Intelligence company with the mission to decrease the time and cost of developing drugs. Lokavant was spun out of Roivant Sciences, a NASDAQ-listed biopharmaceutical company with tens of drugs in development. The team at Lokavant noticed that companies like Roivant – despite investing hundreds of millions into their R&D pipelines – struggle with operational risk in their trials. The best way to mitigate operational risk in clinical trials is to enable clinical teams to be proactive and predictive rather than reactive. That’s where Lokavant and its products – Oversight and Insight – come in. Both products collect, correlate and deliver data insights on clinical trials. Oversight is Lokavant’s risk-based quality management application, alerting study teams and executives of emerging risks at the study-, country-, site- and patient-level. Insight enables sponsor and CRO customers of Lokavant to benchmark the performance of their trials against Lokavant’s proprietary repository with data from >2,000 past trials. Lokavant’s products are easy to use for study teams, deliver machine learning models to generate forward looking insights, and are reliable, secure, and scalable. Lokavant deploys its products on top of its Clinical Intelligence platform, a foundational technology layer that ingests and harmonizes data in a source agnostic and real-time manner to surface actionable insights earlier than ever before.

When we first began developing our products and platform there was a need to find a partner that helped provide the best environment to start building and deploying without bogging down the business with unnecessary costs and effort. Lokavant quickly realized that AWS could help provide the solutions that we urgently needed. Working with AWS has helped Lokavant cut the time spent on infrastructure-related work and provided the time and resources to let the technical teams focus on building Lokavant’s growing suite products. It also has allowed Lokavant to allocate spend where most appropriate to the company’s current maturity. AWS provided the solutions that Lokavant needed to be stable, to scale at the right pace for the company, and to deliver solutions much faster.

Our microservices architecture sits on top of Amazon Elastic Container Service (Amazon ECS), with Snowflake supporting our backend data marts and analytics stores. Amazon ECS, with autoscaling capabilities, allows us to easily deploy containerized business logic and micro-service interfaces exposed through Elastic Load Balancers (ELB) that scale up and down as needed. AWS Fargate descopes our effort, allowing us to focus on container-based solutions without having to worry about our own infrastructure and operating system complexity.

To complement our core backend services running on Amazon ECS and ELB, we distribute our customer-facing applications on Amazon CloudFront and expose composite services on Amazon API Gateways, where we leverage AWS Lambda to quickly deploy business logic that integrates with multiple back-end AWS services. The ability to deploy these user interfaces in multiple regions around the world via Cloudfront provides faster access to our international customers. Lokavant also uses Amazon Cognito to authenticate our application and APIs, and this allows us to seamlessly integrate with our customers who leverage SAML2 and OIDC Single-Sign On federation, which is a core requirement of many sponsors and CROs.

Fig. 1: A high level representation of the application runtime architecture. The application single page React front-end is delivered through Cloudfront and all business logic is maintained in the microservice layer which is deployed either in AWS ECS Fargate or AWS Lambda. Microservice endpoint management is done either through an Elastic Load Balancer or via API Gateway. API end points are secured using oauth integration with AWS Cognito. Application auto-scaling is managed using auto-scaling policies defined in AWS Cloudwatch tied to CPU usage metrics.

On the backend, Lokavant is delivering novel and innovative analytics, not only in static point-in-time insights, but also with predictive analytics. These analytics leverage our ability to mine and correlate data and outcomes from historical clinical studies to build and train models that deliver leading indicators of study risk. Currently, we deliver our model engines in separate containers that provide preparation, ETL and data mart creation, and then deploy models built in a wide variety of Python-based packages that train and execute in parallel. Lokavant drives all data ingestion and data lake management via AWS Managed Apache Airflow, which directly interfaces with Amazon Simple Storage Service (Amazon S3) and Snowflake to prepare and synchronize multiple dependent workflows requiring a high degree of correlation.

With the use of containerized model engines, Lokavant is configuring robust atomic building blocks to work in distributed ECS Fargate, executed through Managed Apache Airflow. These containerized engines also allow us to leverage new capabilities with bring-your-own container integration via notebook-driven interfaces on Amazon SageMaker. These additional tools will help Lokavant run investigative data science in combination with existing model outputs to identify new trends and data patterns that will ultimately enhance our products and improve trial operations for our customers.

Fig. 2: A high-level representation of the architecture supporting data and analytics workflows. Customers deliver data through secure S3 interfaces. Managed Workflows for Apache Airflow is leveraged to create multi-step orchestrations to detect, organize, and load data. The respective DAGS where the workflow orchestration resides is delivered to MWAA through AWS CodePipeline and managed in Github. More complex business logic including data ingestion ETL and data mart ETL, as well as model training occurs in containers that run in AWS ECS Fargate. These containers are also built and deployed through AWS CodePipeline.

In addition to deployment and execution environments, Lokavant has invested heavily in Continuous Integration and Continuous Deployment (CICD). We are constantly improving our stakeholder productivity, and we are creating automated delivery channels for our application development, data engineering and data science groups. Lokavant makes use of the AWS CodePipeline and AWS CodeBuild to integrate our GitHub repos and automate deployment to our auto-scaled ECS and Lambda infrastructure. All our pipeline configurations are driven from AWS Systems Manager parameter stores, and our ability to spin up a completely new vertical stack is measured in hours, not days or weeks. Our build pipelines work interactively through AWS Chatbot and Slack to keep our developers apprised on the status of builds and deploys, as well as any required approvals or gated interactions for our higher-level environments.

Fig 3: A high level representation of the code build and deploy process. Code is sourced from Github at the time of code merge configured within pipelines constructed in AWS CodePipeline. AWS CodeBuild invokes the respective build definition and then the runtime artifacts are deployed into Cloudfront, AWS ECS Fargate or AWS Lambda. All build events are tracked and key notifications are posted to Slack through AWS Chatbot for both informational and “action needed” purposes.

Conclusion

Operational risk ultimately stems from the way data are managed and interpreted in clinical trials. In the past decade, there has been a proliferation in the data that are being generated by clinical trials as well as the e-clinical systems that are being used to collect these data. One would expect that more data are always better, but fragmentation of data across e-clinical systems are pervasive and such legacy solutions have certainly not been designed for the growing influx of data from disruptive innovations like decentralized clinical trials.

Lokavant and its products have helped small and large sponsors and CROs turn this data challenge into a data advantage. In one study, Lokavant improved the quality of the study by saving >12 patients from loss to follow-up, a significant risk for the study that would have cost >$1M. In a separate study, Lokavant prevented a 6-month delay by unearthing site non-compliance issues months before traditional methods for site management would have. Through its continued investment in data science and analytics, Lokavant has also built machine learning models that improve the accuracy of enrollment forecasts by up to 70x – a common source of mistakes during the planning phase of a new trial.

As we continue to scale across thousands of news trials and build new products in clinical intelligence, we will continue to leverage AWS as a core partner in our mission to cut the time and cost of clinical development.

Rohit Nambisan is a product leader with management experience across multiple healthcare technology domains, including Big Pharma, medical devices, personalized medicine, Health IT, health data and analytics, and AI. Prior to leading Lokavant, Rohit was most recently the Head of Digital Product at Roivant and the Head of Product at Prognos.

Brian Monroe has been designing, building, securing, and supporting DevOps and IT infrastructure for over 20 years in multiple industries including B2B SAAS, Financial Services, Retail Pharmacy and now at Lokavant. Brian has a very deep and diverse background in multiple technology domains supporting all facets of the application eco-system.

Welcome Back to the San Francisco AWS Startup Loft

AWS Editorial Team — Wed, 06 Apr 2022 20:26:49 +0000

AWS is excited to announce the re-opening of the San Francisco AWS Startup Loft on April 18, 2022. If you’re a burgeoning startup looking for support, we invite you to immerse yourself in this innovative space and its thriving community at 525 Market Street. The SF AWS Startup Loft is open to all startups building on-cloud and is a free benefit for existing AWS Activate members.

Co-working space for founders, builders, and teams

The SF AWS Startup Loft is a hub where founders, builders, and their teams gain access to the expertise they need to build, launch, and scale their startup.

You’ll find co-working space that supports teams of all sizes, enabling every startup to work in the way that’s best for them: whether you’re the white-boarding type, love a good round-table discussion, or want to focus on heads-down work before bringing it all together.

The SF AWS Startup Loft is where creators can build their MVP, test assumptions and messaging, and remove barriers associated with finding the talent, capital, and networks they need to accelerate. In short, the Loft is a microcosm for innovation, with the tools needed to help your startup move at light speed.

Re-opening is only the beginning

There’s more to the SF AWS Startup Loft than co-working space. Regardless of where you are on your startup journey—whether you’re ideating on the next billion-dollar idea or preparing for your first (or tenth) exit—the Loft welcomes you with an abundance of support and resources. Over the next few weeks, you can take advantage of new programs and tools, including:

Ask an Expert reservations with AWS Solutions Architects and Business Development pros
A podcasting studio for startups to use, to get messaging and content out to followers.
Applied programming through events, Startup Workshops, Labs, and more.
An on-demand video library (Startup Academy) Loft users can explore and learn from.

Book on-site consultations with best-in-class AWS technical experts and business leaders to get tailored advice for your specific stage of development. Dive into a specific problem or challenge with an AWS Solutions Architect in the Loft’s collaboration rooms. Reserve a co-working space and refresh and re-energize with amenities in our state-of-the-art facility. Sign up for Startup Workshops and Labs to unlock immediately applicable insights. In short: come with challenges, and leave with solutions!

Not an AWS Activate Member? Join at the Loft!

If you’re not an Activate member yet, visit our website to sign up ahead of time, or sign up at the door.

We couldn’t be more excited to welcome you back to the San Francisco AWS Startup Loft on April 18, 2022. Make your way to 525 Market Street and come see where innovators gather in San Francisco. We can’t wait to see what you build, launch, and scale.

Video: How AWS Empowered BlockFi to Scale at Their Own Pace

AWS Editorial Team — Tue, 05 Apr 2022 15:39:04 +0000

“We have employees spread across 22 countries,” explains Adam Healy, BlockFi’s Chief Security Officer. “And we’re needing to provide internal services to them, at scale, securely, in a robust way.” BlockFi also had to be able to scale externally, as its customer base grew from an initial 150,000 people to its current count of over 600,000 active clients in just 18 months.

“AWS was a clear market leader,” says Healy. “Not only in their cloud capabilities and infrastructure and platform services, but some of the new and innovative things they’re doing around machine learning and data science.” By partnering with AWS, BlockFi also benefited from being able to access the marketplace and other third-party services that AWS has relationships with, allowing them to solve problems more quickly.

As for BlockFi’s customers, “It’s generally a more tech-forward audience and a more security-aware audience,” says Healy. “They have the confidence that they’re placing in us and through us—organizations like AWS that we’ve chosen to collaborate with—to safeguard their assets and provide them financial services at a speed and at a velocity that they would expect in a highly innovative market.”

If you’re interested in learning more, watch “How BlockFi scales to meet demand for crypto asset services,” BlockFi’s speech from AWS re:Invent 2021.

How Citus Health Uses AWS to Provide Secure and Real-Time Virtual Patient Care

AWS Editorial Team — Tue, 29 Mar 2022 14:05:22 +0000

by Vincent Zheng, AWS Associate Startup Solutions Architect, and Christine Samson, AWS Startup Solutions Architect Manager

Prior to starting Citus Health, Co-Founder and CEO Melissa Kozak spent eight years providing direct care to patients receiving home infusion therapy, in which patients are administered medication intravenously or subcutaneously. Throughout her time as a nurse, she identified a gap in tools available to patients to connect directly with their care providers when they had questions or were experiencing issues. Kozak recalled a moment in which a patient’s IV pump had been beeping all night long; he reached out to an on-call service, where he was transferred three times and then disconnected, resulting in missing his dose. With the help of Citus Health’s Co-Founder Shadid Shah, a thought leader in the digital health care space who had built and deployed multiple electronic health record systems in his career, they were able to bring the idea of patient-centric communication tools and resources to life.

How it works

Citus Health is a digital health transformation startup that enables real-time and secure collaboration between health care teams, partners, and families to provide the best patient experience possible while positively impacting the financial outcome of the care provider. The web and mobile-accessible application provides patients and providers features such as secure two-way communication, centralized nursing notes, documentation, scheduling, and video sessions. Citus Health also allows for third party system integrations where companies can integrate their applications with Citus Health’s microservices running on AWS. For example, specialty pharmaceutical companies can integrate with Citus Health’s platform as a way to send and receive documents and important patient information both from the patients themselves and from health care providers to provide a smoother and easier experience in providing patients with the medications that they need.

How Citus Health utilizes AWS

Figure 1 – Citus Health Architecture Diagram

Citus Health’s application is built using a combination of PHP, NodeJs, Angular, and Ionic which interacts with microservices running in AWS infrastructure through the use of APIs. Their website and microservices are hosted on Amazon Elastic Container Service (Amazon ECS) with the Elastic Compute Cloud (Amazon EC2) launch type. As their application grows, managing the deployment, structure, and scaling of these containers become increasingly complicated. ECS’s container orchestration capabilities aim to simplify configuration options while handling the heavy lifting.

Citus Health utilizes containers for multiple microservices that provide features such as real time messaging, credentialing, and enrollment. Amazon ECS handles the placement of these containers on Amazon EC2 clusters that come pre-installed with Docker. It also simplifies scaling, monitoring, and managing these instances through the AWS management console. Communication between instances these containers reside on and Amazon ECS is done through the ECS Container Agent which also comes pre-installed on the EC2 instances within a cluster. Utilizing the Amazon EC2 launch type provides greater control over details such as the instance family and size that these microservices containers will reside on. Citus Health is able to decide on the instance family, size, and along with other configuration details, allowing them to leverage the benefits of container orchestration, while still maintaining control of specific aspects of infrastructure management that is vital to ensuring their application has the proper resources and capacity that it needs.

Citus Health also utilizes tools such as private subnets, which contain their microservices that handle tasks like credentialing, secure messaging, and enrollment. These API services are placed behind a Network Address Translation (NAT) gateway, which allows for the APIs interact with other AWS resources and the internet while blocking external connections outside of the AWS Cloud. Sitting in a different private subnet are their PostgreSQL databases running on RDS which are used to store data that is generated by their application. Network Access Control Lists (NACLs) are placed in front of all of their subnets to filter out unwanted connections.

Lessons Learned

When Citus Health was first assessing a migration to AWS, they had two major concerns: data sovereignty for compliance with PIPEDA regulations, for one of their customers in Canada and finding the easiest way to migrate from a managed service provider (MSP) Rackspace to AWS. For their data sovereignty concern, Citus Health worked with the account team to determine how to be compliant with PIPEDA. Although the shared responsibility model for compliance and security was not new to Citus Health, they needed more clarification on how PIPEDA specifically works on AWS. They weren’t sure if the HIPAA Business Associate Agreement (BAA) covers PIPEDA, or if either party (Citus Health or AWS) needs to do more in order to be compliant with PIPEDA. Citus Health met with the AWS account team and a compliance specialist to discuss this concern. They learned that the shared responsibility model still applies, that AWS takes care of security and compliance of the cloud while they are responsible for security and compliance in the cloud. They also learned to use AWS Artifact for generating reports related to PIPEDA compliance standards and to get a BAA from AWS. However, they wanted to also know where it states that AWS does not move a customer’s data. Upon collaboration with the account team, a compliance specialist, and Citus Health’s own lawyers that they learned that this is stated in the AWS customer agreement page that they sign.

Citus Health also realized that they need to be cognizant of where the data is going if the data is being sent via SNS to other regions or if the phone containing these messages are taken to a different country. They also need to be able to account for those scenarios and figure out what services to use to ensure their data stays in Canada.

For their migration concern, Rackspace runs on AWS and just manages the infrastructure for their customers. By working with the account team, Citus Health discovered that they don’t need to use the usual ways to migrate their data and servers from Rackspace onto AWS. Since Rackspace runs on AWS, they have an AWS account ID that Rackspace manages for them. By taking the advice of the account team, Citus Health contacted Rackspace directly and asked if the account ownership can be transferred from Rackspace to AWS directly so they wouldn’t need to actually migrate anything between accounts or platforms. Rackspace is able to do this, so for any early stage startup, this can help expedite migrations and alleviate work on the customer side.

Citus & AWS together

Throughout Citus Health’s Journey, the company has learned to utilize a variety of tools and services to best protect both the customers and the businesses they serve; security has to be top of mind at all times. Internally, Citus Health utilizes a separate VPC with a bastion host, which vets and blocks unwanted connections while allowing authorized connections to connect the bastion host. Through the bastion host, their DevOps team and developers can reach the main VPC’s resources through VPC peering. Separation of their resources with two separate VPCs helps isolate their resources and adds an additional layer of protection and separation. Security groups are put in place that act as virtual firewalls for each VPC. Applications can communicate with Citus Health’s microservices through authorized API calls while blocking unwanted connections. NACLs work in conjunction with security groups which adding an additional layer of security. Security groups work at the instance level, preventing connections from reaching the resources within an VPC while NACLs filter connections as they come into the subnets that these resources are placed in. NACLs also allow you to explicitly deny connections while security groups only allow you to implement allow rules. NACLs are also stateless meaning that traffic will be denied unless an allow is explicitly defined. Monitoring tools such as CloudTrail and IAM Analyzer also help in keeping track of who is accessing what resources and allows for Citus Health to keep track of resources and roles that are shared with external entities. All of this has been put in place to protect patients and their information while also ensuring that other health care companies utilizing Citus Health’s services are protected as well.

As Citus Health continues to grow, they need to scale their resources to be able to meet the demands of their customers. After their initial move from being managed by Rackspace to AWS, Citus Health had only set up Amazon EC2 instances along with an Elastic Load Balancer (ELB) to distribute traffic. The capacity that they had spun up was enough to support their current traffic and demand. Citus Health was initially only deployed in the United States and as they started to grow in size they eventually expanded into the Canada region. With this growth came a need to implement autoscaling measures as a way to spin up extra capacity to ensure that their customers continued to receive a smooth and seamless experience. Citus Health also began to leverage Amazon ECS within the Canada region and had broken up their application into microservices as a way to decouple their application and leverage container orchestration.

They also have a multi-AZ setup for their Amazon Relational Database Service (Amazon RDS) instances for disaster recovery. Data is synchronously replicated between the primary instances and if the primary RDS instance fails, Amazon RDS will initiate an automatic failover to a standby instance without any administrative intervention. When it comes to providing patients with a method to communicate with healthcare providers, elasticity ensures a smooth and seamless experience while availability ensures that patients stay connected and have an open line of communication to the resources that they need.

Conclusion

Startups move at a very fast pace and details like security, elasticity, and availability can end up neglected due to wanting to release a product or service as quickly as possible. By utilizing AWS, Citus Health was able to leverage built in tools and services to secure their environment and ensure that their services remain available and resilient. Ensuring that things like security groups, NACLs, load balancers, autoscaling, and multi-AZ setups are properly configured will help keep your application available and secure so that you can best provide your customers with the smoothest experience possible in the long run.

Vincent Zheng is a Startup Solutions Architect at AWS based in New York City. He helps startups across different industries by providing prescriptive guidance and best practices on how to leverage AWS to design and deploy applications at scale. When Vincent is not at work, he enjoys rock climbing, eating out at restaurants, playing video games, going to the movies, and going on walks.

Christine Samson is an AWS Startup Solutions Architect Manager based in New York City. She leads a team of SAs who are trusted technical advisors for startups. She enjoys traveling, exploring new places to eat, playing the piano, and playing sports such as basketball, volleyball, golf, and tennis.

AWS Activate and 11 Other Essential Tools for your Startup’s Digital-first Workplace

AWS Editorial Team — Wed, 23 Mar 2022 12:28:58 +0000

This past February, Slack launched a new Digital Toolkit that brought together 11 must-have tools for your startup. As part of this campaign, AWS Activate – a free program specifically designed for startups and early stage entrepreneurs – teamed up with Slack to provide you with resources and expert support to go to market faster, and lower your startup costs.

By joining AWS Activate through this promotion, you’ll receive:

$1,000 in AWS credits
$350 in AWS Support credits
Best practice guidance
Pre-built AWS infrastructure templates
Exclusive member-only offers
Personalized content, tools, and resources

Slack also partners with these innovators across the industry to build integrations into their solutions, and they are extending those partnerships to make the tools themselves more affordable too. Be sure to take advantage of these limited time offers from:

Slack: Connect your people, tools, customers and partners and free up the way you work. Access the best of Slack with 30% off your first year. Learn more.
Notion: Get 20% off a Notion Team Plan for 12 months. Eligible for companies with 200 employees or less. Learn more.
Loom: Get 30% off a Loom Business plan. Learn more.
Asana: Get 30% off eligible Asana Business plans for the first 12 months. Learn more.
Zapier: Get 100 tasks with Zapier for free per month. Learn more.
Salesforce: Get 50% off a Salesforce Essentials annual plan. Learn more.
DocuSign: Start a 30-day free trial and get 15% off select DocuSign eSignature plans. Subject to applicable terms and conditions. Learn more.
Brex: Get 25,000 Brex points after spending your first $1,000 with a Brex card. Learn more.
Canva: Unlock 60 free days of Canva Pro, giving you access to premium design tools and content. Learn more.
ActiveCampaign: Get 20% off your first year when signing up for an ActiveCampaign annual Plus plan or above. Learn more.

Whether you are just getting started or already building on AWS, learn how joining Activate helps you go further with help from Slack’s Digital Toolkit.

Ten Mistakes Founders Make on AWS, and How to Avoid Them

AWS Editorial Team — Tue, 22 Mar 2022 13:43:41 +0000

by Britton Winterrose, AWS Startup Solutions Architect and Jordan Patapoff, AWS Startup CTO Engagement Lead

For founders, the only thing longer than the hours you work is the list of tasks to be done. Ideally, you’d get to spend your time talking with your customers and implementing improvements to your product. That’s why AWS fuels the growth of startups at all stages by providing reliable, scalable, and inexpensive cloud computing services that allow founders to quickly innovate, experiment, and iterate. The breadth and depth of AWS services make the platform a powerful tool to build fast and focus on your customers — but with that power comes great responsibility.

The AWS Startup Solutions Architecture team exists because we believe that by helping founders in the earliest stages, more successful startups will reach their long-term goals. In our past roles as founders, my colleagues and I experienced the benefits and challenges of building on the cloud first hand. Now as Startup Solutions Architects, we serve as technical advisors to early stage startups in the world’s best accelerators including Y Combinator (YC) and TechStars. In meeting with thousands of startups one-on-one, we’ve witnessed startups leverage AWS successfully, and seen what does and does not work.

We want to help you stay focused on your customers and building features for your products, so we’ve put together a list of the most common mistakes we see founders make on AWS, paired with advice on how to avoid them to save you time and money.

Mistake #1: Not setting up AWS Budgets

No one likes a surprise bill. Learn why every startup should set up a budget and how AWS budgets makes it easy. This is #1 for a reason: it is your safety net and early warning system. If you aren’t sure where to start, start here! Setting budgets and alerts only takes five minutes.

Mistake #2: Not leveraging business support

Time is of the essence for startups who are building. With help from AWS Business support, you don’t have to go it alone. Learn how to use business support to help fix issues and get answers faster.

Mistake #3: Confusing root accounts and IAM users

Having fundamental security guardrails in place early can save trouble, time, and money down the road. Don’t be discouraged if it feels like a steep learning curve. Instead, check out this post on how to improve security using Identity and Access Management (IAM).

Mistake #4: Not setting up MFA

Multi-Factor Authentication (MFA) helps prevent unauthorized access to your account. For startups an account breach can be a fatal mistake. Learn why you need MFA on your account, then go enable it… now.

Mistake #5: Not using Infrastructure as Code (IaC)

Infrastructure as Code (IaC) allows you to manage infrastructure with configuration files instead of doing things manually though the console, making your life easier and saving time if implemented correctly. Learn if now is the right time to adopt IaC for your startup.

Mistake #6: Rolling your own instead of using a managed service / SaaS solution

One great piece of advice from Y Combinator is to “look for a way in which you can accomplish 90% of what you want with only 10% of the work/effort/time”. AWS managed services are there to help do just that and save you time. Learn about and consider: when should startups use a managed service?

Mistake #7: Using Access Keys when IAM roles can be used

Access keys, IAM Roles, IAM Users… which to use and when? It can get confusing, especially if you’re trying to build fast. Earning trust with your customers begins with keeping their data secure. Learn how setting up IAM Users and IAM Roles can help keep your startup secure.

Mistake #8: Not using Amazon CloudFront (CDN)

Did you know Amazon CloudFront can save your startup money and reduce latency for your application? Learn how a content delivery network can keep your application running at scale.

Mistake #9: Leaving unused resources running

Turning off resources you’re not using is the best way to lower your AWS bill, but it is not always clear how to best identify which resources to turn off. Learn how to extend your runway by lowering your cloud bill through resource tagging and automation.

Mistake #10: Storing everything in a relational database

Database selection is one technical decision that will have long-term impacts on your application’s performance and your team’s productivity. It is worth your time to learn about the various database options available on AWS and how to pick the right one before you build.

We all make mistakes

This is not a list of mistakes to obsess over, nor is it exhaustive in its scope. The impact of each mistake varies depending on the stage of your startup, so now may not be the best time for you to focus on fixing each one. As a founder it is important to be aware of how not addressing these issues can affect startup longevity. Then, use your best judgment to prioritize corrections that impact the security of your account or that could be detrimental to your runway. Consider this wisdom from YC’s Essential Startup Advice:

“It turns out that nearly every startup has deep, fundamental issues, even those that will end up being billion dollar companies. Success is not determined by whether you are broken at the beginning, but rather what the founders do about the inevitable problems. Your job as a founder will often seem to be continuously righting a capsized ship. This is normal.” – Geoff Ralston, Michael Seibel

Be informed, take courage, and keep building and fixing things as you go. We’re excited to build with you.

Britton Winterrose is a Startup Solutions Architect with AWS. Prior to joining AWS he founded two companies. At AWS he works with startups that are participating in Y Combinator, the startup accelerator responsible for launching companies like Airbnb, DoorDash, Dropbox, Instacart, and many others.

Jordan Patapoff is a CTO in Residence on the Startup team at AWS. Prior to joining AWS, he co-founded Clyp, a Techstars-backed company that scaled to support millions of creators.

How Emerald Cloud Lab Is Revolutionizing the Laboratory Using AWS

AWS Editorial Team — Wed, 16 Mar 2022 11:36:33 +0000

Guest post by Ben Smith, VP of Engineering and Kevin J. Hou, Scientific Computing Engineer, Emerald Cloud Lab

Emerald Cloud Lab (ECL) provides access to a highly automated laboratory, equipped with over 200 unique pieces of scientific instrumentation, to any scientist with a computer and internet connection. Our wet lab – picture lab benches, hazardous chemicals, lab coats, and safety glasses – has diverse experimental capabilities with an emphasis on biotechnology, and supports a variety of enterprises including drug development, consumer products, and academic research.

Our platform enables scientists to design, execute, analyze and interpret their wet-lab experiments from anywhere in the world. Scientists simply ship their samples – anything from test tubes and cell-culture plates to commercial products – to the ECL and design experimental protocols through a software interface. These experiments are then run exactly as specified in the ECL workflows. Once the experiments are complete, scientists can analyze and interpret their data in the same software interface, and end up with a well-structured, easily navigable and complete record of the experimental execution and results. The ECL offers a myriad of benefits for scientists at universities, large pharmaceutical companies, and startups, such as:

Reduced Capital Expenditures – The majority of the cost of biotech research typically lies in building and operating a laboratory. Individual instruments can take months to purchase and install, and the total cost of necessary equipment can be upwards of $10 million. Additionally, there are ongoing costs associated with ordering consumables, maintaining instruments, and performing qualifications. The Emerald Cloud Lab offers a year of access to a fully-managed laboratory with state-of-the-art instrumentation for less than the upfront cost of a single instrument. This substantially lowers the entry cost for startups wanting to conduct new and innovative research.

Efficiency – Scientific progress is frequently bottle-necked by the sheer complexity of both experimental design and the operations required to execute them. Many scientists today spend 80% or more of their time managing the logistics of science (ordering materials, setting up instruments, waiting for instruments to run, etc.) instead of the science itself (forming hypotheses, designing experiments, analyzing results, etc.). The ECL allows scientists to focus on the science, leaving the logistics to the ECL, with the confidence that their experiments will be performed exactly as specified.

Additionally, the Emerald Cloud Lab uses a combination of laboratory automation and technology-driven operational efficiency to provide experimental throughput unattainable in a traditional lab. The cloud lab operates 24/7, and by taking advantage of these economies of scale, enhances scientific output while preserving low operating costs. All of this has translated into a 5-8x increase in the number of experiments a scientist can perform using the ECL as compared to a traditional laboratory.

Reproducibility – Reproducing experimental results is one of the hallmark challenges of scientific research. Unreproducible results are typically caused by unreliable instrumentation, insufficiently documented protocols, and lost or incomplete data. The Emerald Cloud Lab addresses these issues using heavy integration with technology. For example, we use robotic liquid handlers for precise chemical preparation, and integrated software to ensure detailed measurements (e.g. temperature, weight, volume) are recorded into digital lab notebooks for every operation performed in the laboratory. This combination of automation, extensive monitoring, and procedural operations ensure no data is lost and all protocols are well-documented. Further, experimental protocols are abstracted in our Symbolic Lab Language (SLL), ensuring that repeating an experiment with identical settings is as simple as rerunning a few lines of code.

Why AWS?

As our cloud lab expands to meet growing demand, we have faced a growing need for scalable, on-demand compute. This has been driven by the addition of instruments to the physical lab and by an increase in customer-driven research, where compute-intensive work such as simulations and image-analysis have been integrated into experimental workflows. Serverless, fully-managed AWS services were the perfect solution to address this growing need. The inherent scalability and pay-per-use cost structure of these services provide key advantages for our rapidly growing business, and have enabled us to seamlessly expand to meet growing computational demands.

To address the needs described above, we built Manifold, a microservices-based architecture that runs on AWS Fargate, which we use to provide on-demand asynchronous compute for both internal and external users of the Emerald Cloud Lab. Manifold enables users to run arbitrary, containerized code with full access to our laboratory APIs and database, i.e. with full access to the cloud lab and experimental data.

Since its full deployment in November 2021, Manifold has enabled full automation of routine lab scripts, such as inventory checks, instrument qualifications, and experiment scheduling. Whereas previously we had difficulty reliably running ~1,000 scripts/week with compute resources we had on premises, AWS has allowed us to seamlessly scale to ~5,000 scripts/week in the cloud. This has expanded the scope of automation in the lab, enabling routine tasks to run more frequently with fewer errors. Importantly, migration of our routine lab infrastructure to AWS has also enabled us to achieve 100% uptime on our remote compute services, allowing our computational architecture to remain functional through unpredictable events such as power outages and lab shutdowns.

Manifold architecture

At a high level, Manifold consists of an API which allows users to define compute jobs, and backend AWS components which schedule and run these jobs. Our users interact with the service through desktop and browser apps, from which they make API calls (in the Symbolic Lab Language) to upload job definitions to our internal database, Constellation. To connect these job uploads to the rest of our components, we used Amazon Kinesis to stream changes from Constellation to the rest of our serverless architecture.

Fig. 1: A high-level visualization of the architecture of the Manifold service. Users interact with the service by submitting jobs using APIs in the Symbolic Lab Language. The job definitions are then uploaded to the Constellation database. Using AWS Kinesis, database changes are then streamed to Lambda functions, which provision resources and submit jobs for remote computation on AWS Fargate using Amazon SQS.

Amazon DynamoDB is used to store internal state for the AWS Lambda functions.

Next, we use AWS Lambda functions as a lightweight, scalable method to process the stream of database changes. These lambda functions, in conjunction with DynamoDB tables for storing internal state, are used to provision resources and schedule jobs. The lambda functions then pass compute jobs to our compute service, Fargate, using an SQS queue. Fargate is the keystone of the Manifold architecture – the ease of rapidly deploying containers with varying configurations and permissions has been critical for meeting the diverse computational needs for both our customers and internal developers.

The database-driven architecture described above not only allows for on-demand asynchronous computations, but also allows for jobs to run at scheduled times or in response to other changes in the database. Internally, we use the Manifold service to run routine jobs such as unit testing, sensor checks, and maintenance scripts. This same infrastructure is also used to provide customers with a service where they may submit long-running analyses such as simulations and video analysis to managed resources, with full access and integration with experimental data.

Challenges

Deploying Manifold required us to overcome several challenges arising from the variable load and diverse nature of jobs submitted to the service. Fortunately, AWS provided us with crucial tools for addressing these problems.

In the cloud lab, jobs are often submitted in bursts – for example, after completion of a long experimental protocol, a large number of jobs may be submitted to process the newly generated experimental data. AWS Fargate enables us to effortlessly scale compute up and down to meet these fluctuating demands. More broadly, we have also taken advantage of the monitoring and management tools in AWS, such as AWS CloudTrail and Amazon CloudWatch, to implement rate-limiting, which smooths out usage and enables us to set limits and priority on job submission on a per-user basis.

Following our initial test deployments, we determined that building robust logging into Manifold would be critical towards its success. To provide an optimal user experience, we sought to create tools for users to monitor and manage their Manifold jobs. To accomplish this, we set up infrastructure to selectively upload logging information, status updates, and potential errors from Fargate directly to our Constellation database. This has allowed us to provide users with curated dashboards showing job status, progress, usage limits, as well as error codes and traces for debugging user-submitted computations.

Conclusion

AWS has allowed us to build Manifold, a unified service for remote computing which supports laboratory operations and provides computational services for our customers. The move toward cloud computing at the Emerald Cloud Lab has been obvious and inevitable – in many ways, the ECL is revolutionizing the traditional laboratory analogously to how AWS has revolutionized computing. The ease of use, reliability, and inherent scalability of AWS services has been a natural complement to our cloud lab model, enabling us to deliver an exceptional service to our current customers. Moreover, Manifold will be a key catalyst for future growth, enabling us to meet growing compute demand from customers conducting groundbreaking scientific research, and to support our operations as we increase the size and number of ECL facilities.

	Ben Smith is the VP of Engineering at Emerald Cloud Lab, where he leads the Software Engineering, IT, and Scientific Computing teams.
	Kevin Hou is a Scientific Computing Engineer at Emerald Cloud Lab. His work ranges from designing distributed systems to building analysis software for high-throughput flow cytometry.

The Case for Purpose-Built Databases

AWS Editorial Team — Mon, 14 Mar 2022 13:44:30 +0000

By Cullen Dejean, Sr. Solutions Architect at AWS, and Matthew de Anda, Startup Solutions Architect at AWS.

As an early-stage startup, you have many technical decisions to make while you pursue product-market fit. Some of these technical decisions are reversible, while others are critical junctures with long-term impacts. Picking a database tends to fall into the latter category, which is why a one-size-fits-all approach with a relational database no longer works. You need to take a step back and review the exact use cases you have before jumping into selecting a database. Change your expectation that one database can do everything, and instead choose the database service best fit for the task at hand. AWS offers a broad and deep portfolio of purpose-built databases that support diverse data models and allow you to build data-driven, highly scalable, distributed applications. In this blog post, we’ll cover the factors an early-stage startup should consider when reviewing database options on AWS.

Factors to consider when selecting a database

Experience

Early-stage startups have to focus on building a minimum viable product and demonstrating traction and growth, so any decisions should be made with an eye toward speed to market and available skills. Picking a technology that your team doesn’t have experience with will often increase timelines and result in MVPs that are harder to make changes to in response to user feedback. This is a common reason why using a relational database can continue to be the right approach in the beginning. If your team has experience with a specific data service (e.g. relational, document, etc.) then starting there can be the right choice.

Future scale

Future scale is another factor of which early stage startups must be mindful. Choosing a familiar technology to enable quicker delivery needs to be balanced out by re-analyzing the use case and determining future scaling needs. Being able to leverage solutions that scale automatically with you while continuing to perform as expected can be a force multiplier. It’s always possible to migrate to a different technology later on, but keep in mind that migrations will become more complex as your data grows. Some migrations, like those between Amazon Relational Database Service (Amazon RDS) and Amazon Aurora, are reversible—what we call “two-way door decisions.” Two-way door decisions allow for faster experimentation and embrace a “fail fast” approach. In the next section, we’ll discuss database services such as Amazon RDS and Aurora further.

Purpose-built databases and their use cases

For a long time, relational databases like MySQL and Postgres dominated the database landscape. Now, there are a lot more database types to choose from. To make an informed decision, it’s helpful to assess databases according to their access characteristics and the shape of the data.

Relational databases

A relational database is self-describing, because it enables developers to define the database’s schema, as well as relations and constraints between rows and tables in the database. Developers rely on the functionality of the relational database and not the application code to enforce the schema and preserve the referential integrity of the data within the database. Some typical use cases for a relational database include web and mobile applications. Startups use Amazon RDS and Amazon Aurora for high-performance and scalable applications on AWS. Both RDS and Aurora are fully managed and scalable databases.

NoSQL: Key-value and document databases

As your system grows, large amounts of data are often in the form of key-value data, where a single row maps to a primary key. Key-value databases are highly partitionable and allow horizontal scaling at levels that other types of databases cannot achieve. Use cases such as gaming, ad tech, and IoT lend themselves particularly well to the key-value data model, where the access patterns require low-latency Gets/Puts for known key values.

Amazon DynamoDB is a fully managed, serverless, key-value NoSQL database that delivers single-digit millisecond performance at any scale. For highly performant key-value use cases, we also offer Amazon Keyspaces, a scalable, highly available, and managed Apache Cassandra–compatible database service.

Another relevant database type is a document database. Document databases are intuitive for developers to use, because the data in the application tier is typically represented as a JSON document. Document databases are popular for use cases such as storing and querying content management system data, as well as managing user profiles, preferences, and requests to generate recommendations and enable transactions.

For document data, developers can persist data using the same document model format that they use in their application code, using the flexible schema model of Amazon DocumentDB (with MongoDB compatibility), a fully managed and durable database to achieve developer efficiency and support millions of document reads per second while scaling compute and storage independently.

Data warehouses

When determining the need for data warehouses, it’s important to distinguish between transactional (OLTP) and analytical (OLAP) databases. OLAP databases are larger databases for warehousing and data archiving. For many early stage startups, considering Amazon Athena to handle OLAP use cases can be the right choice. Amazon Athena is a serverless SQL query interface that allows you to analyze data stored in Amazon Simple Storage Service (Amazon S3) with standard SQL.

When your applications begin to require more complex queries and stricter SLAs, this is where the power of data warehouses can help scale your data strategy and requirements. Amazon Redshift is a fully managed, petabyte-scale data warehouse service in the cloud.

In-memory databases

In-memory databases are used for applications that require real-time access to data delivered at microsecond latency. Financial services, ecommerce, web, mobile, and gaming applications have used in-memory databases to build leaderboards, session stores, caching, and real-time analytics. In-memory databases can ease the load off your relational databases, deliver lower-latency results, or replace the relational database and be a primary in-memory key-value data store.

Amazon ElastiCache makes it easy to set up, manage, and scale an in-memory data store or cache environment. Amazon ElastiCache works with both the Redis and Memcached engines. Amazon MemoryDB for Redis is a Redis-compatible, durable, in-memory database service that delivers ultra-fast performance. It is purpose-built for modern applications with microservices architectures. One way of deciding between these offerings depends on whether your use case is ephemeral or requires more durability. Amazon ElastiCache is often used as a stand-alone database, but only for applications that do not require durability. In contrast, MemoryDB is designed to be a primary database. The other factor would be your team’s familiarity with Redis or Memcached engines.

Search

Search databases deliver near real-time analysis and search. Common use cases for search databases include log analytics, real-time application monitoring, and clickstream analytics.

Amazon OpenSearch Service makes it easy for you to perform interactive log analytics, real-time application monitoring, website search, and more. Amazon OpenSearch Service supports OpenSearch and legacy Elasticsearch OSS.

Graph databases

A graph database’s purpose is to make it easy to build and run applications that work with highly connected data sets. Typical use cases for a graph database include social networking, recommendation engines, fraud detection, and knowledge graphs. Deciding if a graph database is the right choice starts with determining whether the data can be best represented as graph structure. Are the breadth and depth of relationships increasing? Queries for relational databases will become slower as the relationships become more complex. Are the models themselves changing enough that schema changes are becoming a burden on your team? Finally, will you need to answer questions about relationships in your data? Graph databases provide this kind of flexibility while delivering complex queries of these relationships. Amazon Neptune is a fast, reliable, and fully-managed graph database service.

Ledger databases

Ledger databases can help track data’s entire lineage. Amazon Quantum Ledger Database (QLDB) is a fully managed ledger database that provides a transparent, immutable, and cryptographically-verifiable transaction log ‎owned by a central trusted authority.

Time series databases

Time series databases handle use cases where applications are dealing with time series data and need to quickly analyze this data using built-in analytic functions. Amazon Timestream is a fast, scalable, and serverless time series database service for IoT and operational applications.

Database Type	Use Cases	AWS Service
Relational	Traditional applications, ecommerce, OLTP transactions	Amazon Aurora Amazon RDS
Key-value	High-traffic web applications, ecommerce systems, gaming applications, financial trading	Amazon DynamoDB Amazon Keyspaces
Document	Content management, catalogs, user profiles	Amazon DocumentDB
OLAP	Traditional data warehousing, multi-tenant BI applications	Amazon Athena Amazon Redshift
In-memory	Caching, session management, gaming leaderboards, geospatial applications	Amazon Elasticache Amazon MemoryDB for Redis
Search	Consolidated logging, personalized search	Amazon OpenSearch
Graph	Fraud detection, social Networking, user profiles	Amazon Neptune
Ledger	Systems of record, supply chain, registrations, banking transactions	Amazon QLDB
Time series	IoT, DevOps, industrial telemetry	Amazon Timestream

Conclusion

As an early-stage startup, one of the more critical decisions you will make is what type of database technology to use. When reviewing the purpose-built databases that AWS offers, start with the use case and define the needed requirements to help filter which database services are a good fit. Then, layer on speed to market and your team’s available skills, but make sure to balance those factors against your future needs. It’s important to spend the time now considering these factors and selecting the database service that’s best for the job.

Hacking for Social Good: How AWS Hypercharged Our Hackathon

AWS Editorial Team — Fri, 11 Mar 2022 14:43:10 +0000

This post was co-authored by Ferdinand von den Eichen, Cofounder and CTO, kineo.ai; and Julia Ostheimer, Machine Learning Scientist, kineo.ai and Team Member, DSSG Berlin.

In recent years, digital technologies have transformed businesses and nonprofit organizations alike. Nonprofits are often especially overwhelmed by the amount of data they accumulate and lack the resources to generate value from it. In response to this challenge, the charitable organization Data Science for Social Good (DSSG) Berlin was founded in 2015 with the mission to enlist volunteer data scientists and analysts to help nonprofits utilize their data properly. With more than 250 volunteers worldwide, the DSSG community supports organizations in social good areas through projects in education, the environment, public health, and more. To bring these volunteers and nonprofits together, DSSG Berlin hosted Datathon; a data science hackathon providing challenges involving each of the participating organizations, which participants solve by hacking throughout the weekend.

Achieving speed and excellence with AWS during a hackathon

Hackathons thrive on the participants’ ability to move at breakneck speed, and Datathon 2021 was no exception. Data from the participating nonprofit organizations was prepared by DSSG’s data ambassadors months before the event, and organized by Datathon 2021 sponsor kineo.ai, Berlin’s Machine Learning (ML) consultancy. Kineo.ai’s mission is to enable German businesses to adopt AI, from the verification of early ideas all the way to production and millions of requests and data points.

The prepared data contained millions of entries on weather, COVID-19, and other factors. Traditional approaches weren’t going to cut it, so the team used Kubeflow Pipelines running on top of Amazon Elastic Kubernetes Service (Amazon EKS), to obtain the data. They followed by using AWS Glue to catalog and partition the data combined with Amazon Athena to provide quick, comfortable ad hoc query capacity.

During the event itself, participants focused on generating insights from these various data points, with Amazon SageMaker Data Wrangler at the heart of the solution. Using its simple, streamlined APIs, candidates accessed portions of data and put it into the tools of their choice, refining as needed. While some participants were content to carry out that kind of exploration on their local machines, others found they needed more horsepower to accelerate model training. Those participants were able to transfer their existing code quickly to Amazon SageMaker Studio Notebooks, enjoying significantly faster speeds, thanks to greater compute and GPU power. Of course, plenty of teams just wanted to use ML, rather than implementing everything from scratch. These teams made great use of some of the AWS high-level ML APIs, such as Amazon Textract, Amazon Comprehend, and Amazon Forecast.

Fig. 1: Participants in the DSSG Datathon 2021, celebrating in the Closing Ceremony.

37 participants. 8 data ambassadors. 6 Datathon volunteers. 3 organizations.

More than 30 participants from Germany and around Europe joined the Datathon from remote locations.

Use Case 1: Investigating influences on Germany’s bike traffic

The German Cyclists Association, the Allgemeine Deutsche Fahrrad-Club (ADFC) e. V., provided the first use case. Here, the main objective was to understand the impact of weather conditions and the COVID-19 pandemic on bike traffic. To improve Germany’s bike infrastructure, participants correlated different data sources to extract meaningful insights to boost the ADFC’s political lobbying efforts.

Fig. 2: Dynamic Time-Series Visualization of the Bike Counting Data in Berlin

The teams applied different data science techniques, from dynamic time-series visualizations as seen in Figure 2, to time-series decomposition techniques (trend, seasonality, residuals), and correlations analysis. One of the main findings was that there is a correlation between rainy weather and the bike traffic during rush hours (correlation coefficient of -0.28), showing how rainy weather negatively impacts the number of people biking.

For Berlin in particular, data showed the increase in bike traffic over the past 4 years, resulting in more bike-related accidents. One team sourced bike accident data and compared it with the location of bike counting stations (Figure 3). Future steps for the ADFC could be to observe the number accidents at certain counting stations to evaluate which ones are more likely to be accident-prone hotspots.

Fig. 3: Heatmap visualizing the occurrences of bike accidents in Berlin (the redder, the higher the number of bike accidents) and the blue dots showing the locations of bike counting stations

The team investigating the influence of the COVID-19 pandemic on bike traffic in Berlin was able to provide insights on the varying levels of bike traffic during separate lockdown instances in Berlin (Figure 4).

Fig. 4: Number of bikers in Berlin counted since COVID-19 pandemic has started across several bike counting stations (black boxes highlight lockdown and 3G restrictions in Berlin).

Use Case 2: Assessing the parking situation in Berlin

With data provided by district office Friedrichshain-Kreuzberg of Berlin, the goal for this use case was to generate insights into the district’s parking situation, through an analysis of OpenStreetMap (a free wiki world map) data. Participants applied geospatial analysis techniques to correlate parking information with demographic data, to visualize areas in terms of their parking spot density.

Fig. 5: OpenStreetMap with an overlay of the density of parking spots per km² in the city of Berlin.

Along with showing some general analytics (Figure 5), illustrating with the density of parking spots per km², participants discovered that even though 9.8% of Germany’s population has severe disabilities, only 0.05% of the parking spots in the district of Berlin Friedrichshain-Kreuzberg are built to accommodate people with disabilities. Figure 6 illustrates the distribution of parking spots for people with disabilities using blue dots.

Fig. 6: Overview of disabled parking spots with an overlay of the density of parking spots per km² in the city of Berlin.

Another participant examined the location data from Point of Interest (POI) locations in Berlin, such as schools, pubs, and tourist attractions, and see how many parking spots are available in the surrounding areas. To illustrate this, a parking consumption ratio was built by dividing the number of counted cars in the vicinity of a POI to the number of counted parking spots (Figure 7).

Fig. 7: Screenshot of a street in Berlin where the Polygon of a 40m distance surrounding a POI is drawn with rectangles indicating cars and points indicating parking spots.

Use Case 3: Analyzing skill shortages at the German Red Cross

Finally, Germany’s national Red Cross organization, the Deutsches Rotes Kreuz e.V., is experiencing a shortage of skilled workers. The main objective of this use case was to build a better visual of the shortage by analyzing data from the German Red Cross job platform. Participants applied natural language processing and data analysis techniques to highlight which jobs would be more likely to attract new applicants and which factors may be driving the skill shortage. Due to privacy regulations, detailed findings from this use case cannot be presented.

Fig. 8: Examples of the Topic modeling for the job platform of German Red Cross showing the highest word-use probabilities.

Conclusion

Datathon 2021 was a great success driven by the desire in the IT and data science community to provide social good. With the seamless integration of the DSSG Berlin and kineo.ai team, and the infrastructure support of AWS, participants provided quantifiable insights for the nonprofit organizations, and many will go on to present their findings at events such as The Future of Care event, hosted by the German Red Cross. AWS services were critical to enhancing and speeding up these efforts that have the potential to impact so many. If you’d like to get involved as a volunteer or want to participate in the 2022 Datathon or help out with nonprofit data projects, reach out to DSSG Berlin.

Ferdinand von den Eichen is Co-Founder and Managing Director at Kineo.ai. He has been building cloud architectures on AWS, Azure and Google for the last 10 years. He is excited to push the boundaries of AI tech and wants to build modern organisations on the basis of personal growth, ownership, trust and empowerment.

Julia Ostheimer is a Machine Learning Scientist at Kineo.ai where she focuses to develop impactful solutions by combining human domain expertise with machine intelligence. In her free time, she volunteers for DSSG Berlin as a team member and was part of the core team organizing the Datathon 2021.

Camino Financial Is Using AI Technology to Loan with Empathy

AWS Editorial Team — Thu, 10 Mar 2022 13:31:33 +0000

In 2014, Sean Salas and his twin brother, Kenneth, co-founded Camino Financial, a financial technology platform that aims to bring affordable credit to under-banked Latinx microbusinesses. “The US Latinx microbusiness economy has an unmet credit demand of $20 billion,” says Sean Salas. “We use the power of AI technology to make sure that capital is more accessible and affordable than anywhere else where they could find that credit.”

The business was inspired by the Salas twins’ mother, who immigrated from Mexico to the United States and opened over 30 restaurants in southern California called El Mexicano.

Unfortunately their mother lost her businesses when the Sean and Kenneth were 12 years old. “As you can imagine, that was a character building moment for all of us, to say the least,” Salas shared.

After receiving MBAs from Harvard Business School, the Salas brothers founded Camino Financial, determined to create a company that used technology to do good in the world and invest in minority-owned businesses. “Financial inclusion was always our north star as a solution to addressing some of the struggles that we saw our mom experience,” says Salas. “That really was the seed that planted Camino Financial.”

Camino Financial focuses on using technology to create an inclusive lending process. This means reaching people who might not be eligible to receive loans from banks. “We’ve created our alternative score to FICO, which we call the ‘Camino Score,’” explains Salas. “Not only are we able to lend to that 25% of our borrower base that don’t have a FICO [score] and otherwise would be credit invisible, but we also help them develop a FICO score.”

In order to build the technology that would power Camino Financial, the Salas brothers focused on creating a strong foundation. They named their backend platform “Luna.AI” after the Luna pyramid in Mexico. “It’s a beautiful pyramid, and as you get closer to the top, you start getting a different feel for the world that you would not otherwise get at the bottom,” explains Salas. “We think about our tech stack as a pyramid. As you get closer to the top, it becomes more proprietary and insightful. But like any good pyramid, you need to have a very strong foundation.”

“When we think about our tech stack and the foundation, it really does start with AWS,” says Salas. He’s especially excited about the ways in which AWS is using its underlying technology to facilitate the development of AI. “Being able to add more bells and whistles enables us to focus on our core and not necessarily over-focus on the foundation and that infrastructure.” The result is money saved, lower-cost loans for members, and the freedom to innovate.

Using the human element to drive AI is a large factor in terms of what makes Camino Financial unique. “When I think about the human factor, I think of empathy,” says Salas. “Without that empathy and understanding of the market, AI can actually work against making capital more accessible and affordable to underserved communities. I don’t think we talk about empathy and its impact on AI enough in the fintech industry,” Salas notes. “I really hope that Camino Financial can lead by example.”

So far, Camino Financial has deployed over 5,000 loans to communities of over $120 million into under-banked microbusinesses. And they’re just getting started. “The US Latino GDP on a standalone basis aggregates to $2.6 trillion,” Salas explains. “That’s the eighth largest economy in the world, and it’s actually growing faster than India and China.” Put simply, the Salas brothers believe that the US economy can’t afford not to invest in these businesses.

Looking forward, Camino Financial plans to keep expanding and offer more credit products to members so that they can build wealth over time. “The engine of economic growth and innovation in the United States has always been our small business economy,” says Salas. “And we fundamentally believe that affordable credit is one critical component of that Camino path—to build that generational wealth.”

Extend Your Runway by Turning Off AWS Resources When Not in Use

AWS Editorial Team — Tue, 08 Mar 2022 14:41:16 +0000

by Guillermo Mansilla, AWS Solutions Architect, and Srijit Mitra, AWS Senior Startup Solutions Architect

Cost optimization is a priority for customers of all sizes, but particularly for early-stage startups. Founders want to make sure their financial resources are being used properly and they’re getting the most out of AWS.

Here are a few approaches that early stage startups can implement easily in order to get the most out of AWS in a cost-effective manner.

Tag your resources

You can assign metadata to your AWS resources in the form of tags. Each tag is a label consisting of a user-defined key and value. Tags can help you manage, identify, organize, search for, and filter resources. AWS tags can be used for many purposes, but we will cover two of them: breaking down costs and setting up automation.

Let’s assume you are running two versions of your platform, alpha and beta. You can assign a tag with the key Version and value “alpha” to all your resources running the alpha version of your platform. Resources include AWS services like Elastic Load Balancing (ELB), Amazon Elastic Compute Cloud (Amazon EC2) instances, Amazon Relational Database Service (Amazon RDS) databases, Amazon Elastic Kubernetes Service (Amazon EKS) clusters, Amazon SageMaker notebooks and endpoints, AWS Lambda functions, etc. Similarly, for the resources running the beta platform, the Version tag will have the value “beta.”

1. Breaking down AWS costs

As an early stage startup, understanding your expenses in order to come up with a pricing model is a priority. AWS Cost Explorer allows you to filter your spend according to User-Defined Cost Allocation Tags. For tags to appear on your billing reports, you must activate your applied tags in the Billing and Cost Management console. It can take up to 24 hours for the tags to appear in the console. Follow the instructions here to activate the tags you created and applied to your resources. In the screenshot below, we are filtering all resources that have a project tag with the value “flask-api.” Breaking down costs in this manner helps you quickly validate pricing models as you test different prototypes.

2. Setting up automation

The AWS Instance Scheduler is an AWS solution that enables customers to configure custom start and stop schedules for their Amazon EC2 and Amazon RDS instances. The solution is easy to deploy and can help reduce operational costs for both development and production environments. Customers who use this solution to run instances during regular business hours can save up to 70% compared to running those instances 24 hours a day.

One thing to keep in mind when using the AWS Instance Scheduler is that you’ll be deploying infrastructure to manage infrastructure. The resources being deployed are all serverless and will scale down to zero when not in use, which is the best cost-optimization technique. We will talk about serverless and their main benefits in the next section. (If you don’t have previous experience with serverless, then this is a good opportunity for you to get started.)

The AWS Instance Scheduler also uses tags to filter which instances you want to control. During the initial configuration of the Instance Scheduler, you define a tag to identify the Amazon EC2 and Amazon RDS instances that you want to stop and start on a schedule you define. When you create a schedule, the name you specify is used as the tag value that identifies the schedule you want to apply to the tagged resource. For example, a user might use the solution’s default tag name “Schedule” and create a schedule called “uk-office-hours.” To identify an instance that will use the “uk-office-hours” schedule, the user adds the Schedule tag key with a value of “uk-office-hours.” Tags can be applied to multiple resources using the AWS Tag Editor.

Adopt serverless, managed services

So far we have shared some best practices to help you get the most out of AWS’ pay-as-you-go model by simply turning off unused resources. But what if there were services that didn’t even need to be turned off? AWS offers a suite of serverless, managed services whose pricing model is based on actual usage rather than run time. For example, say you are using an EC2 instance to host a Node.js RESTful API in production. You are being billed for this instance whether any requests are being served or not. That same web application can be hosted using Amazon API Gateway and an AWS Lambda function where you will only be billed per request. As an added bonus, the first million API calls made to a REST API on API Gateway are free every month. Now add on the benefits of a fully managed service (automated scaling, reduced operational overhead, etc.) and you are saving both time and money by letting AWS take care of the undifferentiated heavy lifting. Here is a step-by-step guide on setting up your first serverless application.

Conclusion

Tagging your resources will give you better visibility into your spend, and setting up automation makes it easy to turn off resources you are not using. Moreover, adopting serverless technologies can lower the cost of hosting your workloads—as well as the expenses incurred in time and effort of managing and maintaining resources.

This is by no means an exhaustive list. Rather, it’s just the starting point of the various cost-optimization options available on AWS. As your startup matures, we recommend you take a look at this article for additional ideas, and to keep your momentum going, consider applying for AWS Activate, a startup-focused program designed to accelerate long-term growth by issuing AWS credits and providing architecture guidance and access to subject-matter experts. These benefits are designed to give you the right mix of tools, resources, and expert support so that you can succeed with AWS while optimizing performance, managing risk, and keeping costs under control.

Have fun, and build on!

Guillermo Mansilla is a Solutions Architect based in Orlando, Florida. He has helped startups in the USA and Canada build and architect their applications on AWS. Before joining AWS, Guillermo spent over a decade working as a Software Developer. When he’s not at his desk, you can find him playing chess tournaments at his local chess club.

Srijit Mitra is a Sr. Startup Solutions Architect based in Toronto, Canada. He works with early stage and emerging startups to help bring their disruptive ideas to fruition. He has spent over a decade building and implementing DevOps practices at organizations of all sizes and transforming their software development practices.

AWS Healthcare Accelerator in the UK Announces 12 Startups Selected for the Inaugural Programme

AWS Editorial Team — Mon, 07 Mar 2022 13:15:49 +0000

The first AWS Healthcare Accelerator in the UK cohort represents a variety of healthcare solutions backed by diverse leadership, with 50% of the cohort identifying as women-owned or minority-owned startups.

We are excited to announce the selection of the 12 participants for the first ever AWS Healthcare Accelerator in the UK. These 12 companies were selected from 117 applications from 13 countries, including the UK, by a panel from Amazon Web Services (AWS) and PUBLIC. PUBLIC is a leading government-focused technology company that brings a strong network of healthcare startups and an international presence to the AWS UK Healthcare Accelerator experience.

Announced in October 2021, the AWS UK Healthcare Accelerator in the UK is a four-week programme that cultivates and promotes innovative startup solutions that achieve the quadruple aim of improved patient experience, improved clinician experience, better health outcomes, and lower cost of care.

The 12 participants in the programme are working to solve many of the strategic imperatives of AWS healthcare customers in the UK.

Dignio

Dignio connects patients and healthcare professionals through a digital platform that disrupts the traditional way of delivering health services and opens up new possibilities for a more sustainable solution. The award-winning solution, enables safe and cost-effective care, whilst at the same time liberating clinicians from routine tasks.

Sapien Health

Sapien Health is the world’s first digital clinic for surgery, helping people prepare their mind and body for surgery through sustainable lifestyle changes.

Dr Julian

Dr Julian is a mental healthcare platform that aims to increase accessibility and regulation of mental healthcare provision connecting the patients to vetted, online trained clinical psychologists, CBT therapists and counsellors via our secure video/audio/messaging link from their phone, tablet, or computer.

WYSA

WYSA provides mental health support for everyone by helping users self-manage stressors by blending AI-guided listening with professional expert support. Services are anonymous, available 24/7, and clinically safe.

DDM Health

DDM Health is a multi-award-winning provider of digital therapeutics that improve patient health outcomes and unlock clinical efficiency.

PEP Health

PEP Health uses AI to provide deep insight into what patients are thinking in real time. Our proprietary Natural Language Processing technology transforms millions of patient comments into measurable & comparable data so that they can be used to improve experiences, drive better outcomes and democratise the patient voice.

Remedy Rx

The future of healthcare has patients at the centre of the ecosystem. Remedy’s platform captures the 95% of data that sits outside of the healthcare system by creating a direct, dynamic and longitudinal link between patients and their physicians leading to life changing results. Better data. Better care. Better trials.

Birdie

Birdie delivers the most comprehensive, built-as-one technology platform for home care providers. Through the integration of data into a single source, they helps home care providers of all sizes offer a superior level of service to older adults in the UK – personalised around their needs, their health and their happiness. Birdie calls this Personalised Home Care.

Infinity Health

Infinity is a SaaS task management solution for healthcare, providing staff with tools to plan and coordinate care, monitor performance, and make real-time operational decisions that improve safety and efficiency. Infinity replaces outdated technologies and inefficient processes, seamlessly integrating with existing systems to automate and optimise high-volume, mission critical tasks. The activity data collected by Infinity can also be used to build predictive models that inform strategic resourcing and planning decisions.

Abtrace

Abtrace supports healthcare systems to transform the way long-term health conditions are detected, monitored and treated. The platform that acts as data driven layer sitting on top of EHRs prompting proactive interventions and earlier diagnosis.

C the Signs

C the Signs is a clinical system that uses AI mapped with the latest evidence to identify patients at risk of cancer at the earliest and most curable stage of the disease.

Thymia

Thymia analyses speech, video and behavioural data gathered via specially designed video games to help clinicians assess mental health conditions objectively and accurately, starting with depression. The company’s software aids clinicians before, during and after each appointment empowering them to find the right diagnosis and treatment faster, improving patient care, all whilst saving time and costs.

Over four weeks, the 12 selected startups will receive AWS Promotional Credits, specialised AWS training, mentoring from healthcare domain and technical subject matter experts, business development, go-to-market guidance, and investment guidance. The companies will also have potential proof-of-concept opportunities with public sector healthcare customers.

The programme brings in healthcare industry leaders to collaborate with startups on topics ranging from business models, regulatory pathways, clinical validation, electronic health record integration, and more. The programme also fosters collaboration opportunities with AWS customers and AWS Partner Network (APN) Partners looking for innovative healthcare solutions.

The accelerator programme culminates in a virtual demo day on April 25, 2022. Demo day is a showcase event featuring presentations from the startups and AWS leaders. Sign up to watch the showcase here.

Applications were judged on several factors, including the innovative and unique nature of the project, the overall value the solution will bring to the healthcare industry, the creative application of AWS to solve problems, and the team’s ability to deliver on an identified opportunity.

Learn more about the 12 AWS Healthcare Accelerator finalists and discover how AWS for Health supports customers and partners. See how PUBLIC helps healthcare startups to scale and succeed.

Dr Rowland Illing is the Chief Medical Officer and Director of International Public Sector Health for Amazon Web Services (AWS). He has responsibility for healthcare strategy and operations for AWS internationally, excluding the US & China. This encompasses healthcare providers, payors and health technology companies. He is passionate about the delivery of person-centred care, increasing access and improving outcomes at a lower cost by accelerating the digitization and utilization of healthcare data.

Rowland is an Honorary Associate Professor at University College London, who trained in both Cambridge and Oxford Universities. He is a Member of the Royal College of Surgeons of England, Fellow of the Royal College of Radiologists and Senior Fellow of the Faculty of Medical Leadership and Management in the UK. He continues to present and publish on topics including digital healthcare and AI, is a member of the Digital Health Roster of Experts for the WHO and member of the European AI Alliance for the European Commission. LinkedIn, Google Scholar

Helping People Quit Smoking Through Financial Rewards

AWS Editorial Team — Fri, 04 Mar 2022 16:32:18 +0000

Guest post by Hadi Javeed, CTO &; Co-Founder, Vincere Health

Smoking is still the leading cause of preventable death, and the pandemic made matters worse: cigarette sales increased in 2020 for the first time in over 20 years. There are over 34 million adult smokers in the US, many of whom are lower income. Reaching them and providing effective support at this scale while keeping costs down can only be achieved with the aid of technology and carefully designed patient experience.

Vincere Health is one of few health technology platforms built for people in diverse socio-economic categories. Unlike traditional cessation programs, Vincere offers low-cost access to addiction healthcare using reward-based habit training. Our belief is clinicians in the loop are integral to lasting behavior change, and the technology serves to facilitate and personalize this relationship at scale.

Need for a personalized “quit journey”

Smoking cessation is best represented as a journey, not a single event. The personalized care experience that is individualized to each participant’s needs has shown increased rates of success conquering addiction with care compliance rates up to 80% and quit rates up to 35%. Participants are encouraged to set compelling yet reachable weekly goals using our program and stay connected to their health coaches for support.

Our platform allows health coaches to create a personalized participant quit journey.

Health coaches can create a care journey for multiple days/months that can track multiple events just like on a calendar, e.g, breath test at a specific time/day, complete a self-reported outcome survey, or participant/coach video call appointment.
Coaches can define financial reward criteria for achieving a certain goal. e.g, participants can earn rewards for completing a self-reported survey, breath test habit compliance, or making it to the appointments.
Define personalized reminders and notifications. These reminders could be configured to go out at different intervals looking at the user information for behavioral reinforcement and motivational nudges.

During the program, we collect and aggregate lots of data from participants, to measure engagement based on habit compliance, take proactive actions, and inform health coaches about participant triggers or relapses.

Using AWS Cloud to scale

We chose AWS as our cloud provider because AWS provided the necessary tools to help us build a HIPAA-compliant platform. The plug-and-play nature of AWS architecture helped us iterate faster as a startup with a small team.

Let’s explore the main components behind our architecture.

We embrace microservice architecture that runs on top of AWS Fargate. Fargate helps us a lot to avoid the operational burdens of managing servers and allows us to scale, meeting the growing needs of our workloads.
Amazon Relational Database Service (Amazon RDS) is our main transactional database. Amazon ElastiCache complements our core database in performance by caching read-heavy data.
We use Amazon Redshift as the central data warehouse and Amazon S3 as a scalable data lake, merging all application data with events, engagement data, device data, and rewards for analytics and understanding our participants better.
Amazon Chime SDK helped us build audio/video communication tools faster, which are leveraged by our health coaches to communicate and keep a personal connection with participants.

Going Forward

Due to our personalized smoking cessation platform built on AWS and our amazing health coaches, we have been able to launch successful clinical validations with leading institutions such as Boston Medical Center and their research has been accepted by a prestigious American Thoracic Society journal.

Our strategic partners and individual participants are seeing higher program satisfaction rates of 82% Net Promoter Score, 68% reduction in tobacco usage across the population when CO (Carbon Monoxide) was measured objectively using the devices, and higher program engagement rates where on average 7.4 weekly touchpoints were measured during the clinical trial.

AWS proved to be the best architecture for Vincere Health because it provides capabilities to allow us to comply with HIPAA with ease and confidence. Achieving our goal started with data gathering and learning more from participants. Our next steps are to continue to build our platform and add more intelligent layers on the platform using Machine Learning and advanced data gathering techniques to scale our program and reach out to participants proactively who are in the need of the most.

Learn how to deploy a reference architecture for HIPAA in the AWS Cloud.

Hadi Javeed, CTO & Co-Founder, Vincere Health

Hadi has many years of experience in developing and architecting enterprise-grade applications. He finds a deeper passion for creating impact in health care through building software. His expertise focuses on technical direction and providing strategic ways of bridging ideas to engineering execution.

How Polly’s Curated Biomedical Molecular Data Streamlines MLOps for Drug Discovery

AWS Editorial Team — Wed, 02 Mar 2022 14:19:19 +0000

By Swetabh Pathak, Elucidata founder, and Sahil Rai, Engineering Manager at Elucidata

The application of machine learning to biomedical sciences is crucial to attain success in drug discovery. Biomedical data is large and unstructured, rendering it difficult for ML applications. Equally, many human diseases involve several variables that change dynamically and differ from patient to patient. Understanding this complexity requires unlocking large-scale heterogeneous biomedical and patient data. Machine learning, therefore, can help bioinformatics scientists and drug researchers identify meaningful patterns from large datasets that can lead to better odds of success in the clinic.

Advancements in technologies like next-generation sequencing and mass spectrometry have led to an abundance of data. However, the utilization of this data is still challenging as most of it is not well-curated. The state-of-the-art machine learning models that get better with more data require a viable level of data quality. In sharp contrast, today, most of the biomedical molecular data is available in semi-structured and unstructured formats. When developing new drugs for novel diseases, a faulty model can lead to completely off-the-mark outcomes which are only understood much later in the process. This is where platforms such as Elucidata’s Polly come in handy.

Polly is a DataOps platform that allows data scientists to access ML-ready data generated from data repositories, proprietary experiments, and publications. Curating data into a high-quality resource is the core value proposition of Polly. Polly complements the burgeoning data-centric AI movement that seeks to employ data as a tool to improve ML outcomes.

Using Polly, scientists can manage the life cycle of biomedical molecular data end to end. Polly’s capabilities include:

Curating data from proprietary, public, and premium sources using active learning to ensure that data is machine-actionable and analysis-ready.
Storing large volumes of heterogeneous data in one central location.
Leveraging powerful querying capabilities across dataset, sample, and feature level metadata through SQL.
Analyzing and visualizing data with the help of a scalable computational platform.
A toolbox that allows users to either utilize pre-existing bioinformatics pipelines or create custom ones.
Sharing analysis and results with internal and external collaborators.

Use cases of biomedical molecular data are atypical of other domains where machine learning is used. Data is typically large in volume and low in velocity. Formats are heterogeneous depending upon the kind of data. There are few common metadata standards and crucial metadata is human-generated. Search and discovery of data is hard when data is present in multiple sources, is siloed, and is not labeled consistently. A consistent querying interface is often not available. Polly solves these challenges by incorporating DataOps methodologies. Powered by AWS, Polly’s infrastructure is fast, secure, and scales seamlessly.

How Polly works

AWS has a robust lineup of services that serve a platform’s DataOps demands such as building and managing data pipelines, ML pipelines, and hosting applications at scale with great ease. Four critical workstreams are deployed on AWS:

Ingestion & ETL infrastructure
Storage & querying infrastructure
Data curation infrastructure
Scalable compute infrastructure

Workstream 1 – Ingestion & ETL infrastructure

Polly ingests and processes hundreds of thousands of datasets every week with an infrastructure capable of scaling to millions of datasets. Polly’s infrastructure is designed to be fault-tolerant and cost-effective. Polly’s ETL infrastructure uses Amazon EC2 Spot instances to reduce the cost of data ingestion by almost 90% compared to using regular on-demand instances. The steps involved in Polly’s ETL pipelines are:

Datasets from multiple sources are added to an Amazon Simple Storage Service (Amazon S3) bucket. There is an S3-to-lambda trigger set up on every dataset that carries out the upload/delete/update function. Next, the dataset ID is sent to the Controller lambda.
The Controller lambda adds a row in the Amazon DynamoDB table and assigns a job ID. The status of the job is marked as “in-queue”. DynamoDB is used to store Job ID, job metadata, and job state. (Note: Each dataset update is considered as one job).
After adding an item in the DynamoDB, the Controller lambda sends a message to Amazon Simple Queue Service (Amazon SQS). The message contains the job ID.
Messages in the SQS trigger the Scaling lambda.
Scaling lambda starts scaling out the ECS Cluster (spot fleet), which fetches the job metadata based on the job ID.
These ECS clusters transform and load the datasets to Amazon OpenSearch and in S3.
Once loaded, the status of jobs is updated on the DynamoDB table (as Success/Errored).
The time-based Amazon CloudWatch event rule keeps triggering the scaling lambda. Scaling lambda keeps on checking “Approximate number of messages in the queue”. In case of no messages, scaling lambda starts scaling in the spot instances. The last three steps are repeated until there are no messages in the queue.

Workstream 2 – Storage & querying infrastructure

Once the datasets are ingested, the goal is to be able to query and search them effectively. A dataset has two major components: metadata and data matrices. The schema of metadata is semi-structured. On the other hand, data matrices are structured entities, mainly containing numerical values. Search is facilitated by the metadata of the datasets. Some searching algorithms also utilize the underlying data matrices. Hence metadata and data matrices are stored in the Amazon OpenSearch service and Amazon S3 bucket (as Delta Lake) respectively. One challenge with this approach is that it is difficult to query two different sources separately and then combine the results. That’s where Amazon Athena comes into play.

Amazon Athena lets federated queries run on two different sources using SQL syntax, delivering a unified interface to our users. To do this, we have used the following strategy:

The ETL infrastructure loads datasets into Amazon Elasticsearch and S3 bucket. The data in S3 is stored as a Delta Lake.
Elasticsearch is used to store datasets with diverse metadata such as types of diseases, drugs, organisms, etc.
Delta Lake is used to store large 2D data matrices, which contain numerical data. Delta Lake enables us to store and query on-disk data using ACID transactions.
Amazon Athena is then used to query both Elasticsearch and S3 bucket. Querying Elasticsearch is facilitated by the ES-Athena connector. A Glue data catalog is used to store the schema of the data present in Delta Lake.
API Gateway is used in combination with AWS Lambda to proxy the Athena endpoint. Having a custom API layer over Athena ensures that users have consistent access to all Polly’s services.

Workstream 3 – Data curation infrastructure

The metadata quality of a dataset is improved by adding additional fields and filling the gaps in existing fields. This is done by our proprietary ML framework called PollyBERT. PollyBERT has been trained on millions of labels generated by our team of experts who label the datasets on our expert curation platform. The training data is used to scale the annotations to millions of datasets using multiple ML models.

Operating this ML pipeline can be a daunting task. This is why we have designed and developed an end-to-end machine learning pipeline using AWS services. This pipeline consists of 3 sequential steps starting with data collection and preparation, followed by model training, and finally deployment and hosting of the inference model using Sagemaker.

Polly’s data labeling infrastructure has 3 major components:

1) Collecting Training Data

Data labeling happens on Polly’s expert curation infra by expert annotators. For example, an annotator might label a dataset with the associated disease or drug.
After receiving the labeled data, the data is pre-processed using Amazon ECS clusters and stored on Amazon S3 Bucket.

2) Model Training

Labeled data obtained from step 1 is used to train supervised machine learning models. Training happens on Amazon ECS clusters using a custom training algorithm. Training algorithms are dockerized and the image is stored in Amazon Elastic Container Registry (Amazon ECR).
The trained parameters which are obtained from model training are stored in an S3 bucket.

3) Model deployment

Model deployment and inference take place on Amazon SageMaker.
To ensure consistent authentication with all the Polly services, we don’t expose SageMaker endpoints directly to the users. Instead, users can interact with these ML models through Polly APIs and SDKs.

Workstream 4 – Scaling computational workloads

There are many ways our users consume data on Polly. It can be through web apps, R-based Shiny apps, command-line jobs, or Jupyter notebooks. Self-hosting these applications can become challenging as there are many aspects that need to be managed, like scalability, availability, deployments, and collaboration. Polly solves these challenges by providing a managed solution that does the undifferentiated heavy lifting for the user. For hosting applications, users can directly upload docker images on the platform. For using Ipython notebooks, they can create a new notebook using Polly UI. Additionally, users can choose the size of the machine for applications, jobs, and Jupyter notebooks.

For enabling scalable computational applications:

Polly uses Amazon ECR to store applications and workflows in the form of dockerized images.
Polly extensively uses Amazon Elastic Kubernetes Service (Amazon EKS) to deploy applications. The orchestration and scaling happen on Amazon EKS.
Jupyter hub is integrated with EKS to deploy and scale Jupyter applications and there is a custom proxy layer that is used to scale the rest of the dockerized application on EKS which internally consumes EKS API.
A combination of AWS Lambda and API gateway is used for the API interface.
Users just have to upload a docker image to Amazon ECR using their Polly credentials. The shared instances/pods are used to allocate resources.

Elucidata’s goal has been to double the data on Polly every quarter. Hence, it’s pertinent to create data and machine learning pipelines needing minimal human intervention so that the increase in data does not necessitate an increase in team size. When the rate of ingestion can vary from zero bytes on a particular day to TBs on another day, it’s crucial to be able to scale seamlessly. With AWS as a partner, we have been able to utilize the elasticity of the cloud effectively; and adapt to whatever the data may throw at us.

Learn more about Amazon Athena, an interactive query service that makes it easy to analyze data in Amazon S3 using standard SQL. Athena is serverless, so there is no infrastructure to manage, and you pay only for the queries that you run.

Prior to starting Elucidata, Swetabh Pathak was a founding member and product manager at start-ups in technology and affordable STEM education, and has an integrated Masters degree in Mathematics and Computer Science from IIT Delhi.

Sahil Rai is an Engineering Manager at Elucidata. He leads the data infrastructure team managing critical data pipelines and storage infrastructure. He has extensive experience architecting and building solutions on AWS.

To Deliver Rich Content, Startups Need the Right Infrastructure

AWS Editorial Team — Tue, 01 Mar 2022 20:45:09 +0000

by Vinay Arora, AWS Startup Solutions Architect

Startups face enormous challenges when it comes to delivering content to consumers. Costs are a constant worry with the challenging task of getting to market before funding runs out, and reaching target audiences online with the right content at the right time is a major undertaking. In order to compete in today’s visual-first world, startups must deliver rich content — music, videos, and graphics — quickly and flawlessly across all touch points, providing seamless and compelling user experiences.

There are other potential pitfalls, too. To generate real value, startups must be lean and agile, prioritizing high-value functions and testing them with users throughout the development process. Early prototyping can allow startups to detect and correct faults, but identifying problems is pointless if companies can’t act quickly to address them. To stay nimble, startups need an infrastructure and services that will allow them to run their workloads rapidly.

Getting the right infrastructure and services in place is particularly important as startups grow. In a startup’s early stages, when speed is a priority, security can often be something of an afterthought — a fatal oversight. Startups, particularly in the financial and healthcare sectors, must show that they can safeguard their data and that of their customers. To protect themselves and maintain the trust of their users, startups must be able to detect fraudulent traffic, enable privacy requirements, fend off Distributed Denial of Service, assaults, and handle the phasing out of third-party cookies. Put simply: Implementing security best practices can mean the difference between success and failure.

Finally, working at a global scale only adds additional complications. Startups that want to build a global website or offer content to people around the world quickly and efficiently need the right kind of infrastructure to support their goals. Luckily, AWS can help you address all of these challenges.

How a Content Delivery Network can help you work at scale

A Content Delivery Network (CDN) distributes content from an “origin” server — which could be an Amazon Simple Storage Service (Amazon S3) bucket, HTTP server, or Amazon Elastic Compute Cloud (Amazon EC2) instance — to a vast network of cache servers spread throughout the world, and automatically sends end-users to the closest cache server location for viewing content. This results in lower latency for your app and a better user experience. The content can be static, like an HTML page, or dynamic, like an API backend. By utilizing a CDN, you can increase the scalability and performance of your applications.

Amazon CloudFront, Amazon’s global content delivery network, offers the largest infrastructure footprint of any cloud provider — one that’s expanding continuously to help customers deliver better user experiences. It’s been optimized for performance and scalability, security protections are integrated and customizable, and real-time reporting allows you to monitor your application’s performance. Customers have complete control over the service and can make changes on the fly.

Here’s how CloudFront can help you deliver rich content and better experiences to users:

Thanks to CloudFront, data is delivered globally with the highest possible performance. Distributed caches known as CloudFront Regional Edge Caches operate as a caching layer between the edge location and the origin. Previously, when there were no Regional Edge Caches, CDN edge locations were forced to return to the origin when they lacked content. With Regional Edge Caches, however, the edge locations are sent to the origin only if the Regional Edge Caches lack material. This helps minimize demand on the origin, enables scaling of the CDN without scaling the origin, and doesn’t add to your cost. Regional Edge Caches also have larger caches than edge locations, which means objects will remain in them for a longer period of time.

To configure CloudFront, you need to build a CloudFront distribution, which tells CloudFront where to source content. First, you choose the origin servers for your items (S3 buckets or HTTP servers) and upload files, which may include web pages, photos, and media assets. Then, you construct distributions to inform CloudFront which origin servers to use to obtain the files. Additionally, you can decide whether to log all requests and enable the distribution as it is created. CloudFront assigns your distribution a domain name, which you can view in the console, and it will also send the distribution’s configuration to all of its edge locations.

In addition to optimizing performance, CloudFront also helps startups address a number of other challenges.

Free data transfers between AWS cloud services and Amazon CloudFront for origin fetches help your startup stay lean. If AWS origins such as Amazon S3, Amazon EC2, or Elastic Load Balancing are used, there is no charge for data transferred from the origin to CloudFront edge locations (this type of data transfer is known as an origin fetch). And for some origins, like S3, CloudFront can help lower overall data transfer egress costs, which can help extend your startup’s runway.

Moreover, there are pricing options for every level of usage. Pay-as-you-go pricing is simple, straightforward, and flexible, with no upfront fees. If you anticipate a certain monthly spend, the CloudFront Security Savings Bundle could help you save up to 30% on your bill. And for customers who make certain minimum traffic commitments (typically 10 TB/month or higher), we offer custom pricing.

Security is also built into the CloudFront infrastructure, giving startups robust privacy and encryption capabilities. All CloudFront users get the automated safeguards of AWS Shield Standard. CloudFront can also work in tandem with the AWS Web Application Firewall and AWS Shield Advanced to help protect your apps from a variety of sophisticated threats, including DDoS attacks. SSL/TLS is used to deliver secure APIs and websites, where SSL settings are enabled automatically by default. You can easily create custom SSL certificates with AWS Certificate Manager (ACM) and associate them with your CloudFront distributions at no additional cost. ACM also automates certificate renewal, removing the overhead and costs associated with manual renewal. At AWS, security is our top priority, and these features make it easy to implement best practices and keep your data safe.

The best way to deliver rich content and better experiences to users

Slow-loading apps and websites are no longer forgiven in the modern-day startup landscape. Thankfully, Amazon CloudFront provides a content delivery network that allows you to stay agile while delivering new products and updates. It’s designed to provide lower latency and faster data transfer (as compared to serving all traffic directly from the origin server). Developers can take advantage of a full-featured API to create, configure, and maintain CloudFront distributions, and they also have access to tools such as AWS CloudFormation and AWS SDKs to configure and deploy their workloads. Real-time logs, containing detailed information about viewer requests, make it easy for you to monitor the performance of your content delivery and respond quickly to operational events.

Moreover, CloudFront is built on top of the growing AWS infrastructure, enabling connections with all major access networks for optimal performance. CloudFront edge locations are connected to the AWS regions through the AWS network backbone – fully redundant, multiple 100GbE parallel fiber that circles the globe and links with tens of thousands of networks for improved origin fetches and dynamic content acceleration. All told, CloudFront uses more than 275 points of presence in more than 90 cities across 47 countries.

It’s a truly global network that will allow your startup to reach people all over the world, providing rich content quickly and seamlessly.

Learn more about securely delivering content with low latency and high transfer speeds with CloudFront.

Grow Your Startup with Free Tools and Resources from AWS Activate

AWS Editorial Team — Thu, 24 Feb 2022 15:02:16 +0000

This year, over 100 million founders around the globe will embark on their journey to solve the world’s pressing challenges with big, innovative ideas, and AWS is committed to making it faster, easier, and more cost effective to startup and succeed. The potential for these entrepreneurs to disrupt industries and tackle real problems has never been greater, so the ability to build and scale with minimal cost and maximum support is paramount. With AWS Activate, our global startup program, startups can leverage the same technology that successful companies of all sizes are enjoying to build and scale their business. Whether you are just getting started or have already started building, AWS Activate provides you with access to free AWS credits, technical support, cloud experts, pre-built infrastructure templates, and so much more.

AWS credits

With AWS Activate, startups receive up to $100,000 in AWS credits to spend on AWS services. These credits are applied to your account to offset eligible AWS Services’ fees and charges. This means you can quickly get started on AWS at no cost. Startups of every type, size, and industry are using AWS cloud for a wide variety of use cases, such as Machine Learning (ML), artificial intelligence (AI), software development and testing, big data analytics, and customer-facing web applications.

AWS Support plan credits

Along with credits for AWS services, AWS Activate also provides up to $10,000 in AWS Support plan credits so that you have access to AWS experts when you need it. AWS Support engineers save time for you and your team by helping you to move faster in the cloud and focus on your core business. Our cloud support engineers address requests that range from answering best practices questions, guidance on configuration, all the way to break-fix and problem resolution.

Build on AWS templates

Even with a team of developers, you don’t have to start from scratch when building your infrastructure. Build on AWS is a collection of pre-built infrastructure templates and reference architectures developed by experts at AWS and based on AWS best practices. You can deploy these in under 60 minutes simplifying the steps of building reliable, secure, and optimized infrastructure. Build on AWS templates are available from the Activate Console.

Free and discounted startup tools, products and services from leading organizations

We’ve partnered with over 100 trusted organization including Brex, Notion, Slack, and HubSpot to help you get ahead faster. Offers range from free credits to discounted services, products, and memberships that you don’t want to miss.

AWS Activate is just one of the many programs at AWS designed to help startups succeed. In addition to above benefits, you can also take advantage of the AWS Startup Lofts located in San Francisco, New York, Tokyo, Tel-Aviv, and Paris. At the AWS Lofts you can attend in-person or virtual events, hands-on labs, workshops, meet with investors, get answers to your technical questions and collaborate in the co-working spaces.

Ready to join AWS Activate?

Visit www.aws.com/activate to learn more about the program and sign up today!

How Setting Up IAM Users and IAM Roles Can Help Keep Your Startup Secure

AWS Editorial Team — Tue, 22 Feb 2022 11:18:49 +0000

by Kevin Shaffer-Morrison, AWS Startup Solutions Architect and Faisal Farooq, AWS Startup Solutions Architect

For startups, speed is critical — you’re trying to move as quickly as possible to build your product and find product-market fit. Even as you focus on growing your business, however, it’s important not to ignore your security posture. If attackers gain access to your systems, they can disrupt your business operations and cause serious — even fatal — reputational damage. There are some simple steps you can take, however, that will allow you to move quickly while still safeguarding your data. A key one is to employ Identity and Access Management (IAM) users and roles.

An IAM user is an individual, such as an engineer, who needs access to your AWS account, whether through the AWS Management Console, command line interface, or your favorite AWS SDK. Occasionally, there may also be circumstances in which certain third-party applications also need access to your AWS resources (e.g., a CI/CD tool such as CircleCI, or Github actions). In these situations, you can also create an IAM user for that application, with policies for the service to carry out specific actions, such as deploying a new AWS Lambda function or pushing a container image to Amazon Elastic Container Registry (Amazon ECR) and deploying an Amazon Elastic Container Service (Amazon ECS) Task. IAM users can be assigned a variety of credentials, such as passwords and access keys. Creating individual IAM users also allows you to limit their permissions to just the tasks and resources that are needed for the job.

Consider the following scenario: Bob has an AWS account that he needs to access, and he is using a third-party CI/CD system that needs to upload Docker images to the ECR upon every build. Bob can access the AWS account via the AWS web console using his username and password as an IAM user. He also can leverage the AWS Command Line Interface (AWS CLI) and AWS SDKs by using his user access keys. But the CI/CD system isn’t a human, so how can Bob provide it with access to Amazon ECR? To do this, Bob can create an IAM user for the CI/CD system with programmatic access, and an IAM policy that allows writes to ECR. He can then securely share those keys with the CI/CD system so that it can interact with AWS services.

Whether IAM users represent humans or applications, it’s important to follow these best practices:

Use strong password policies when allowing for console access.
Rotate all credentials regularly (both passwords and access keys).
Do not share your access keys or embed them within unencrypted code.
Follow least privilege principles with all users to protect your infrastructure.

There is one potential downside of using IAM users, however: Establishing long-term static credentials can pose a security risk. If an attacker compromises Bob’s IAM user credentials, they can use those credentials for as long as they remain valid — which is even more problematic if the associated permissions do not follow the principle of least privilege. Until the unauthorized use is detected and access is revoked, the attacker can continue using the access keys unhindered.

One way to reduce this threat is to use temporary credentials instead of long-term static credentials. Luckily, AWS provides an easy way to create such temporary credentials: IAM roles.

IAM Roles

An IAM role is similar to an IAM user: It is an AWS identity with certain permission policies that dictate what the role can and cannot do. A role does not have long-term static credentials such as a password or access keys associated with it. Instead, individuals and third-party applications “assume a role” in order to obtain temporary credentials and permissions for the duration of that role, known as a role session. You can read more about IAM roles in the AWS IAM user guide.

This scenario illustrates how roles can help you improve your AWS security posture:

Yuri, a developer working at a startup, wants to access their AWS Staging account to test their changes in a staging environment before the application is deployed in production. Yuri has user credentials to access the startup’s Dev AWS account. Now, when Yuri wants to access resources in the Staging AWS account, she has the permission to assume a Tester IAM role, which grants her privileges to access resources in the Staging AWS account. When Yuri assumes a role, she gets temporary credentials by way of an access key pair for the duration of the role session.

Let’s also consider the permissions of a senior developer on the team, Joy. She also has user credentials in the Dev AWS account, and just like Yuri she has permissions to assume a Tester role in the Staging AWS account. As a senior engineer, Joy also has permissions to assume an IAM role to access resources in the Prod AWS account. Joy’s access to the Prod AWS account is through an IAM role, which provides her with temporary security credentials for the duration of the IAM role session.

This setup ensures that this startup’s AWS accounts are segmented (Dev/Staging/Prod), and only temporary access is granted to the Prod and Staging accounts via IAM roles.

It is important to leverage IAM roles and segment workloads in different AWS accounts. Such segmentation ensures that a misconfiguration or compromised account has a reduced impact. If a user’s credentials are compromised, the attackers will have limited access to other environments.

Applications and services can also assume IAM roles. We often see developers embed IAM user access keys to configure the AWS SDK in an application running on an Amazon Elastic Compute Cloud (Amazon EC2) instance. This is a bad practice, however, because long-term static credentials pose a security risk. If an attacker compromises the EC2 instance, they can access data in Amazon Simple Storage Service (Amazon S3) and other resources based on the permissions associated with the IAM user.

The AWS-recommended best practice is to leverage IAM roles instead. Amazon EC2 instances can assume IAM roles, called instance profiles, to access AWS services, such as Amazon S3. This provides temporary credentials for AWS services and resources that are valid for the duration of the role session.

Once you’ve set up IAM users and IAM roles, there are a few more simple steps you can take to maintain a strong security posture as you build your startup. First, download the credential report from the IAM console and remove any user access keys that are being used in your applications where an IAM role would be more appropriate. Next, run a scan on your codebase to make sure no hard-coded user access keys are still lingering.

From there, you can consider using tools such as Prowler to scan your AWS account and ensure you are following security best practices. You can even send your results to AWS Security Hub for easy analysis. At AWS, security is our top priority, and we’re happy to answer any questions you might have and offer additional guidance.

We know that security can often become an afterthought when you’re moving at warp speed. But implementing these basic security best practices early on will reduce the risk to your startup and enable you to scale your development team more easily, setting you up for long-term success.

When Should Startups Use a Managed Service?

AWS Editorial Team — Thu, 17 Feb 2022 19:59:43 +0000

by Nicolas Menciere, AWS Startup Solutions Architect

A common mistake we see founders make when starting on AWS is trying to implement their own solution when a managed service could be used instead. In the context of AWS, a managed service is a service that is built, maintained, and operated by our highly skilled experts. This stands in contrast to a solution that you deploy onto your own server, where you are responsible for ongoing management and operations — a situation sometimes referred to as “rolling your own.”

We often see developers use open-source software and tools to build their proof of concept locally. Once it’s time to host their application online or scale, however, they either need to refactor part of the application to take advantage of cloud services or keep the application as-is, which often incurs more operational work and higher costs.

Before choosing your tech stack, you must evaluate its ability to scale with your business. You don’t want to build a minimum viable product, get some traction, and then have to rebuild your application once it needs to scale. So ask yourself: Will your architecture still work technically and financially if you succeed?

Choosing the best approach

When selecting the architecture for your application, consider the following questions. Weighing these factors can help you decide whether it’s best to use managed services or implement your own solution.

Operational work: How much time and work is required to maintain your stack? Think about repetitive tasks like version upgrades, security patching, hardware updates, and software deployments. Will these tasks take up a significant amount of time for you and your team? Remember, time not spent on operational work can be spent on creating value for your product.
Infrastructure costs early and at scale: How much will it cost your company to run the stacks at different stages of growth? This is very important, as it often defines how much capital you will need to raise and how profitable your business can become. Cash is king, so the leaner your company is the better. Investors are looking for companies that leverage existing services and focus on creating value.
Complexity: If you choose a specific stack, how complex will it become over time? You want to keep complexity as low as possible, so you can move faster and reduce risk. The more technologies, vendor products, and software you onboard in your product, the bigger and more specialized your IT staff will need to be, potentially driving costs up and increasing overall operational work.
Building features vs. infrastructure: Does using a particular service help you build features faster for your customers? Product market fit often stems from iterating through features, not from having a perfect architecture. Our goal at AWS is to help you get back to writing code that delivers value to your end users.
Agility: How fast can you implement a new feature, and how quickly can you pivot if necessary? Agility is the main value proposition of the cloud and managed services. You can shift gears quickly and innovate faster using managed services.
Reliability: You certainly need to make sure you have as few errors as possible in your application, but maintaining that level of consistency and reliability can be challenging. With managed services, our highly skilled experts focus on implementing and operating the services, so you don’t have to.
Availability: Does your application always need to be up and running? Being highly available is not easy. If your data center, an availability zone, or a region goes down, what happens to your application? Managed services are highly available by default and operated by AWS so you can have peace of mind.
Security: How robust will your security posture be if you roll your own service? Security should be a top priority for your business from day one, but it is a hard problem. Security work creates a lot of overhead in the development process and can slow down innovation. At AWS, security is our number one priority. We have dedicated teams of highly skilled security engineers who work with product teams to make sure our managed services are as secure as possible.
Time to market: How quickly can you release new features? Using managed services means you don’t need to reinvent the wheel and can instead stay lean and focus on adding value to your product.
Portability: Will you need to move your application on premise or to another cloud provider? In this case, you may want to consider rolling your own service. However, you can still use managed services like Amazon Elastic Kubernetes Service (Amazon EKS), which helps you run containers on AWS so you don’t have to manage the Kubernetes cluster yourself. Containers can then be easily ported to another Kubernetes cluster on another platform.
Customization: Do you need a high level of customization to meet your business needs? If so, then rolling your own service may be advantageous, because you can tailor it exactly to your needs. If such customization is not key to your success, however, we recommend that you consider using managed services.

Advantages of managed services

Before you start to write code or deploy open-source software onto a server, you should be aware of the options available to you, so that you can start building the right way, move fast to beat the market, and create a product that will scale along with your business. It pays to shift away from a “build first” mentality and do the upfront research to determine what’s out there, before you commit to a path forward.

AWS provides managed services for the majority of common use cases, so it’s likely that you may be able to leverage an existing solution for your application. As a rule, we recommend that you use a managed service if there’s one available that meets your needs.

AWS managed services also provide a high level of reliability and availability, because we put a significant amount of time, energy, and resources into the services we offer. Taking advantage of these services, rather than deploying them on your own, can save you a lot of the heavy lifting, give you peace of mind, and allow you to focus on creating value for your business instead of dealing with operational work. The AWS pricing model is also designed so that you pay only for what you use, offering you a lot of freedom and flexibility. This allows you to avoid large upfront capital expenses early on, which could otherwise be fatal to your startup.

Advantages of rolling your own

As noted earlier, rolling your own solution allows for a higher level of customization and control. However, we recommend that you consider this approach only if it provides clear upsides for your core business. Will designing and implementing your own solution yield an ongoing competitive advantage?

Regulations and portability are also factors to consider, as you may need your application to use specific software or be hosted on premise or in a specific region of the world. In such cases, you may want to control and operate your stack, and not rely so much on external providers.

You should also factor in your team’s existing internal knowledge, which can speed up initial development. If your team is very experienced with a specific architecture, for example, it may be more efficient to build your application this way. Still, you must consider not just the initial build, but also the long-term costs of operational work and maintenance, as this will make up the bulk of your total costs.

Conclusion

To start from a place of success, consider a cloud-native approach as you start your project. While there are no crystal balls, think of where your application will be hosted, which managed services are already at your disposal, and how it will run at scale. Then consider the programming languages, framework, software, and tools you will need to build it. Embracing this approach will help you make the best decision for your initial deployment.

Video: How Ergatta Partnered with AWS to Develop and Launch the Future of Game-based Fitness

AWS Editorial Team — Tue, 15 Feb 2022 17:17:28 +0000

Enter the Ergatta Rower, a game-based connected rower designed to feel more like playing a sport than attending a gym class. Composed of a rowing machine and a screen made to fit unobtrusively in your living room, the machine is sleek and visually appealing while allowing customers to connect to a broader community.

Like many startups before them, the team behind the Ergatta Rower turned to AWS to help them develop and launch their product. “We knew that AWS would be catering to companies of all life cycles—from early iteration to alpha to beta to launching and scaling,” says Swaminathan. “We knew that there would be help to make sure that we were looking at the right products, looking at the right services, and building things in the right ways.”

Watch the video above to learn more about the co-founders’ journey, including how they’re utilizing AWS’s guidance, services, tools, and technology to build the future of game-based fitness.

Should Startups Use Infrastructure as Code (IaC)?

AWS Editorial Team — Thu, 10 Feb 2022 13:44:15 +0000

When you’re an early stage startup, nothing seems more important than being quick. You iterate fast to find product-market fit. You race to deliver features before your competitors do and at times, trade best practices for quick fixes. Sometimes, it pays to move fast and shore up your technical infrastructure later. However, you can’t escape the fact that you have to spend a little bit of engineering time on foundational work in order to be able to handle growth. Skipping foundational work can end up being a huge mistake, and there’s one in particular that gets made all too often in the startup world: not using Infrastructure as Code (IaC).

We have too often seen startups opt for “easier,” manual solutions that are tedious and not reproducible. Only the individual who creates the solution has an understanding of its configurations. This leads to configuration drifts — an environment where running your workloads in your infrastructure changes over time due to manual changes or other updates. If there is no adept documentation, drifts can cause your organization time and stress, not to mention monetary loss, even over minor issues. It’s important for your customers to have the highest uptime and all their expectations met. Delaying the implementation of IaC makes it more challenging to deliver new features and fixes to users, and it will also take much longer to scale your resources as your user base grows.

In this blog post, we want to highlight why utilizing IaC is so important and beneficial, and show you how useful implementing IaC with the AWS Cloud Development Kit (AWS CDK) can be.

What is Infrastructure as Code?

The “Infrastructure” in IaC refers to all the servers and services that you use to build, test, deploy, and run your app or product. If your app runs on Elastic Compute Cloud (EC2) instances, then the infrastructure includes not only the EC2 instances but also the Virtual Private Cloud (VPC) in which the instances run, the NAT Gateways that your application use to connect to the internet, the Elastic Load Balancers (ELB) that handle traffic from your customers, and the Relational Database Service (RDS) databases in which your application stores data. Additionally, if your application relies on managed queues like SQS or workflow orchestrators like Step Functions, those application-level services are also included in our definition of Infrastructure. Furthermore, if you run your applications as serverless functions or in containers, then the API Gateways and container definitions are also part of your Infrastructure.

The “Code” part of IaC is a script or program that can be executed by an automated process that creates, configures, deploys, and tests whatever infrastructure it has created. In some cases, that could be declarative code, like CloudFormation’s JSON or YAML that gets executed inside the CloudFormation service. In other cases, the code could be the AWS SDK’s API that you could call from many different procedural languages like Python, TypeScript, or Java to provision infrastructure.

A Bit About IaC Tools

While we encourage its use, AWS CDK is not the only tool for IaC. There are several popular IaC frameworks and tools that customers can use with AWS, such as HashiCorp’s Terraform. Many of these tools, Terraform included, work by making API calls to AWS via the AWS SDK. Dependency and state management is the responsibility of the tool.

By contrast, AWS CDK works by turning procedural code written in conventional programming languages into CloudFormation’s declarative template. Beneficially, CloudFormation will maintain the state and dependencies of your infrastructure so that you don’t have to store or manage it. With AWS CDK, you can use familiar and powerful software engineering tools and best practices, such as type-safety, reusability, unit testing, source control branching and merging, and continuous deployment. Similar to other tools, AWS CDK also provides a command-line utility for deploying the generated declarative code to CloudFormation. During this step, AWS CDK can generate least-privilege security roles and policies needed to define and operate your infrastructure and services.

How AWS CDK Helps

Now that we’ve established what AWS CDK can do, let’s talk about a real-world scenario where IaC makes a difference.

Imagine the following situation: Your startup develops an amazing software-as-a-service (SaaS) application targeted at the financial services industry. You’ve done everything right so far — found your product-market fit, developed a scalable and multi-tenant application that makes it simple for new customers to get started with little more than a credit card and a web browser. Early on, though, you decided to save a little bit of time by deploying your application’s infrastructure without IaC. So far, that hasn’t been a problem because you used EC2 auto-scaling to handle growth. As your business grows, larger customers come knocking. Soon, you catch the eye of a global financial services company, and they approach you with interest. If you were to sign them as a customer, it would completely change the trajectory of your startup, so you’re incredibly keen to get them on board.

There is a catch, however. They want you to deploy your application for them as a single-tenant, and in a standalone AWS account that they own. Having to create a new environment from scratch will have an immediate and significant impact on your business. The global financial services company will take longer to onboard, and they will have to wait longer for future infrastructure changes to be deployed. Not only that, but chances are you’ll be updating your multi-tenant SaaS application more often than your single-tenant customer. Over time, your single-tenant customer will notice that they aren’t receiving version updates as often as their multi-tenant counterparts. By having to deploy infrastructure without the benefit of IaC, you’re giving your customers a bad experience. And from there the challenges you face will continue to compound. If you want to grow, you’ll have to take on other single-tenant customers. Each one will add complexity and risk.

When you decided to skip IaC, the key benefit that you’ve missed is automation. Without defining your infrastructure in code, you can’t automate its creation. Instead, you have to manually create infrastructure when you need to deploy a new environment. As your startup grows, you will need to deploy multiple environments. You will also need to maintain the environments that you create, so updating them will also be a manual, risky, and time-consuming process. It might seem like the number of environments will be small, early on. You might have a single production environment and a single test environment. But in reality you are likely to have many, many more environments — and you need to make sure they are reliably identical, or else correct behavior in one environment will not predict correct behavior in another environment.

The Goal

If your goal is to build a modern company using today’s development best practices — which it should be — then you will have environments for developers, unit testing, integration testing, pre-production testing, and production itself. If you have single-tenant environments, you may even have the latter three environments duplicated for each tenant. Provisioning and updating the infrastructure in all these environments manually will be exceedingly difficult, adding time, complexity, and cost to your startup’s operations. And it will only become more challenging as your startup grows.

In contrast, with AWS CDK you can put your infrastructure, application code, and configuration all in one place, ensuring that at every milestone you have a complete, cloud-deployable system. This makes AWS CDK the fastest way to get started with IaC in the early phase of a startup, because developers don’t need to learn a new tool or programming language as they would with CloudFormation’s JSON/YAML complexity or Hashicorp Configuration Language (HCL) with Terraform. AWS CDK equips them with the languages they already know, while allowing them to start learning by making small changes that will ultimately translate into big, scalable changes. It gives you the ability to write less and do more, creating quick but long-term impact and saving you time — now and in the future.

Drive Value with Automation

When we think of all the steps needed to deploy new versions of an application, being able to deploy with automation is key to delivering value to customers. As a startup, you won’t just be creating new environments when you onboard new customers; you’ll need to iterate quickly on your product and gather feedback from your customers constantly. Deploying updates safely with native testing capabilities, without the potential for human error, is one of the key benefits of automation via IaC for startups.

Time to get started! We recommend the AWS CDK Intro Workshop as a fast on-ramp for learning at startup speed. If you’ve postponed an investment of time and resources into infrastructure as code and automation, now would be a great time to make it a priority.

Ed Epstein is the principal at Metapoint Informatics. He lives just outside Vancouver, Canada.

Improving the Sustainability of Crop Farming Using a Data-driven Approach

AWS Editorial Team — Wed, 09 Feb 2022 15:24:42 +0000

Guest post by Dzmitry Yablonski, CTO and Co-Founder of GeoPard Agriculture

Due to the effects of advancing climate change and a growing global population, agricultural professionals are under more pressure than ever before to produce higher yields with fewer inputs. On one hand, the world’s population is expected to increase by 2 billion in the next 30 years according to United Nations statistics, whilst at the same time water, land, labor, and other resources for food production are expected to become scarcer. According to the US Environmental Protection Agency, 24% of all Greenhouse Gas (GHG) emissions originate from agriculture, and action to reduce emission is becoming more and more critical. Meanwhile, COVID-19 remains a challenge. Collaboration between Ag Tech companies is as essential as ever to speed up the delivery of the technologies to the crop growers.

Let’s learn how Geopard, an independent precision agriculture platform, collaborates with Corteva, a leading global agricultural input company, to enhance the capabilities of their Granular Link App.

The Business Case

As a major agricultural company, Corteva Agriscience understands the challenges facing humanity such as climate change, preserving biodiversity, and meeting the increasing demand for food. Sustainable crop farming and biodiversity are integrated into the strategic vision. In Europe, they are continuously increasing the suite of digital solutions to improve the efficiency of agronomic decisions via more smart usage of agricultural inputs.

This year, Corteva Agriscience Europe launched a new product – Granular Link with a tailored Variable Rate Application (VRA) maps. This app augments the company’s physical products (seed and crop protection) with smart recommendations integrated into a decision support tool. That represents an integrated set of digital ag features at the field level to reduce risk and maximize yields. Granular Link provides:

Best-in-class satellite imagery;
Variable-rate maps and integration with machinery;
Optimal harvest date prediction;
Crop disease prediction;
Water needs recommendations;
Photo observation and field notes;
Hyper-local field weather forecast;
Computer vision disease identification;
And a collaboration platform at field level for farmers and Corteva advisors.

From the first day of Granular Link, Flavio Cozzoli, Head of Digital Agronomy and Innovation at Corteva Agriscience Europe, and his team faced multiple challenges that needed to be overcome for a successful commercial rollout. Data sources are inconsistent with each other, and the amount of available data is constantly growing to reach hundreds of terabytes. They required an engine to generate consistent VRA maps at the scale of millions of hectares at the load peaks in minutes. There is also the challenge of the agricultural timeline, where skipping a single crop season could affect the annual plan to validate technologies.

This is where GeoPard Agriculture stepped in to help and collaborate with Corteva Agriscience by providing high-quality, powerful, and scalable data analytics. GeoPard founder’s reputation preceded them: ex-Zoner and a precision agriculture web-based software acquired by Bayer Crop Science in 2015, ex-Xarvio and first internal software engineers/technical managers there.

The Granular Link solution combines the GeoPard Agriculture geoanalytics capabilities, unique knowledge about own products and best-suitable agronomic practices, and the secure scalable infrastructure from AWS to deliver the state-of-the-art technology to growers in the EU.

Our Approach

An agricultural field is a complex multifactorial ecosystem. It includes a variable soil profile with its micro- and macro-nutrients, physical and chemical properties, weather conditions, crop varieties, detailed topography profile, and executed agricultural operations. A single data layer is never enough to accurately describe the crop growth processes. Therefore, GeoPard collects multiple, various data layers and produces insights out of them in an automatically scaled manner over millions of hectares. In the end, the resulting knowledge improves crop production practices, integrates precision agriculture, and enables remote crop monitoring and the transition to data-driven and sustainable agriculture.

The range of available data sources is continuously growing. At the moment, data is produced by:

Satellites, manned aircraft, drones taking into consideration the physical characteristics of installed sensors;
Machinery adopted for precise applying of agricultural inputs (Variable Rate Technology);
High-density scanners;
Laboratory soil samples.

Harmonizing data sources enables further Artificial Intelligence (AI) modeling and human interpretation in an effective way. An understanding of the specific farming systems is essential. The GeoPard team understands the differences inside countries and regions exist. And therefore we treat each of them as a unique entity with its own established farming processes.

Scalability and maturity are inherent characteristics of GeoPard’s offerings. From day one, the team has focused on the cost-efficient and auto-scalable utilization of AWS cloud computing capabilities.

Customer service is not only tailoring the solutions to specific requested agronomic rules; it also includes incorporation into regular processes and flexible integration into existing digital solutions.

Data Sources

Data Sources are the starting point for building any data platform. They are various by nature and include satellite imagery, topography, high-density scanners, and machinery-logged operational data.

Originally, Geopard started with its automated satellite imagery pipeline. The primary focus initially was on multispectral imagery from Landsat4-8 and Sentinel2 missions with historical archives since 1988. Understanding the past vegetation and correlating it to yield, weather and soils provides an excellent tool to understand field variability and predict yield. Extending the satellite imagery suite with Planet, the largest earth observation satellite network delivering a near-daily global dataset with high-resolution data, was the next straightforward step. Meanwhile, the suite is constantly growing and now it is expanding with Synthetic Aperture Radar (SAR) imagery from the Sentinel1 mission.

The comprehensive profile of the field landscape can explain crop vegetation patterns. This can be collected either using remote sensing or machinery-logged GPS data. Among remote sensing sources, there are Aster and SRTM missions and high-resolution LIDAR.

Data collected from sensors on ground machinery mirrors the accuracy of applied agricultural inputs. It represents details of harvesting, spraying, planting/seeding, and fertilizing operations executed in the field. The range of high-density soil scanners is also expanding. Soil sensors can measure much more than electrical conductivity and soil moisture, but also collect data about micro and macro-nutrients. That’s how the soil scanners are becoming the alternative to classic soil sampling probes. The major benefits are continuous field coverage (comparing to selected locations) and scanning frequency (easier to have up-to-date soil conditions).

Data flows are continuous as GeoPard receives data every day. However, raw data is also not enough to make reasonable data-driven decisions. Additional processing is required to produce useful and meaningful results. Extracting knowledge and insights from the acquired data at a large scale is a major challenge for the agriculture industry.

A Powerhouse for Agricultural Data

Technically, the GeoPard powerhouse contains 3 main components: data processing pipelines, data lake, and secured interfaces. It processes the raw original data to produce harmonized datasets and enables both artificial intelligence and agriculture professionals to undertake efficient data exploration and analysis.

The data processing pipelines run in an auto-scalable manner. As many computing resources as needed are rented from AWS during load peaks, and then they are released back once the processing is finished. Artificial Intelligence (AI) and Machine Learning (ML) algorithms are incorporated into the execution of data pipelines.

Data storage is based on the AWS Data Lake approach which stores the data and allows scalable access to it. There are no limitations linked to data size and simultaneous read/write capacity. The storage easily supports terabytes of the data for simultaneous saving (write) and extraction (read) operations. In addition, capabilities of on-demand transformations across various formats are available to simplify integration with any solution.

Secured interfaces for interaction with GeoPard are built on top of Amazon API Gateway, AWS AppSync, and Amazon Cognito. It enables seamless communication with internal and external applications and platforms.

End Products

The GeoPards End Products are part of a greater data-driven revolution in agriculture. They enable the easier/faster integration of precision agriculture technologies. In turn, the precision ag speeds up the transition to sustainable crop farming.

Precision agriculture is a farm management concept based on observing, measuring, and responding to inter and intra agricultural field variability in crops. In particular, field benchmarking helps track the seasonal performance of crops, measure the historical heterogeneity and productivity, and prioritize in-season activities such as harvesting or spraying chemicals.

Smart scouting and sampling assist with the correct detection of locations to visit and to execute the needed measurements and collect samples. Variable Rate Application maps enable the application of the exact amount of inputs to the exact locations where the inputs are required- and only to those locations.

Rich biodiversity and ecosystem balance can be achieved via the use of optimized crop rotations with increasing carbon sequestration and decreasing GHG emissions.

All of these features are available on a large scale measurable over millions of hectares. From day one, we at GeoPard recognized the intensively growing amount of coming data over the next decade. And the cloud foundation based on the most effective AWS approaches (like Data Lake and State Machine) was established to afford excellent scaling capabilities.

Conclusion

By 2050 the world will need to feed an additional 2 billion people using fewer inputs than at present in a more sustainable manner. In terms of climate change, the clock is ticking, placing pressure on the agricultural industry to take immediate action. To make a successful transformation, collaboration between companies is vital. Creating and maintaining synergies between data and cloud computing providers like AWS, precision ag software companies, agricultural input corporations, combined with solutions-based and consulting services, is the most promising avenue for a fast transformation.

Dzmitry Yablonski is CTO & Co-founder at GeoPard Agriculture, former Bayer CropScience/Xarvio Tech Manager, and engineering digital solutions in Precision Agriculture during the past 10+ years. In his spare time he loves playing basketball.

Why Early Stage Startups Need to Use Multi-factor Authentication

AWS Editorial Team — Tue, 01 Feb 2022 16:05:15 +0000

by Xuan Gao, Solutions Architect at AWS on the Startups team

Every startup is unique – there’s no universal to-do list that fits every company. But no matter the nature of your business, security is always of primary importance and should be one of the very first things you address. After all, for most startups, your most valuable asset is your data — your ideas, workloads, and applications — and you need to protect it. We often tell startups to make lots of mistakes, just not fatal ones. A security breach can be a fatal one.

While there is no 100% guaranteed way to protect your data, the one thing you absolutely need to do is use multi-factor authentication (MFA) to protect your accounts. MFA is an authentication method that requires the user to provide two or more verification factors to gain access to a resource. Startups sometimes move so quickly that they don’t take the time to set up MFA, but we highly recommend configuring MFA to protect AWS resources if you haven’t done so already.

The Risk of NOT setting up MFA

Vulnerability: A single password is not enough, regardless of how complex it is. Hackers have ways to crack passwords.
Non-compliance: MFA may be a required component of compliance standards.
Your business reputation: You don’t want to have to answer questions from your customers about why you don’t have sufficient security protocols in place to protect both your business and their personal data.

MFA on AWS

MFA provides an extra layer of security, because it requires users to provide unique authentication from an AWS-supported MFA mechanism in addition to their regular sign-in credentials (username and password) when they access AWS websites or services, and it forces users not to share passwords. Customers can enable MFA for their AWS account root user and IAM users. When you enable MFA for the root user, it affects only the root user credentials. IAM users in the account are distinct identities with their own credentials, and each identity has its own MFA configuration.

Types of MFA

AWS supports three types of MFA devices: virtual MFA devices, U2F security keys, and hardware MFA devices. Virtual MFA devices are software-based apps, usually running on a mobile device, that generate secure, one-time authentication codes that are used as part of the sign-on process. U2F security keys and hardware MFA devices are physical devices that are required to gain access to the accounts to which they are attached. These physical devices are considered the most secure options for MFA, and they can be stored in a safe or lockbox for additional security. Though virtual MFA devices may be more convenient since they can run on mobile phones and are good option in most instances, they are considered less secure than physical devices. At the very least, we recommend that you use a virtual MFA device while waiting for hardware purchase approval or while you wait for your hardware to arrive.

You can also consult the AWS Identity and Access Management user guide for more information on using MFA.

Since MFA is so important, AWS will also provide an MFA device at no cost to qualified account holders. You will be able to use MFA devices to safely access multiple AWS accounts, as well as other token-enabled applications. You can read more about our security initiatives here.

Try it yourself

If you have already configured multi-factor authentication across your startup, congratulations, you’re ahead of the curve! However, if your use of MFA has gaps, now is the time to set it up.

By following these steps, you can set up MFA in just a few minutes.

First, sign in to the AWS Management Console and open the IAM console. Choose a user, then choose the Security credentials tab. Click the Manage button next to the Assigned MFA device.

In the Manage MFA Device wizard, choose the type of multi-factor device. In this example we will show you how to setup a virtual MFA device. You can install a mobile app that is compliant with RFC 6238, a standards-based TOTP (time-based one-time password) algorithm, such as Authy, Duo Mobile, and LastPass Authenticator. These apps generate a six-digit authentication code.

Click Continue, and on next page IAM generates and displays configuration information for the virtual MFA device, including a QR code graphic. The graphic is a representation of the “secret configuration key” that is available for manual entry on devices that do not support QR codes.

Determine whether the MFA app supports QR codes, and then do one of the following:

From the wizard, choose Show QR code, and then use the app to scan the QR code. For example, you might choose the camera icon or choose an option similar to scan code, and then use the device’s camera to scan the code.
Choose Show secret key, and then type the secret key into your MFA app.

When you are finished, the virtual MFA device starts generating one-time passwords. In the Manage MFA Device wizard, in the MFA code 1 box, type the one-time password that currently appears in the virtual MFA device. You then need to wait up to 30 seconds for the device to generate a new one-time password. Type the second one-time password into the MFA code 2 box. Click Assign MFA.

You’re all set! You can see the assigned MFA device information under user Security credentials tab. If you’re using a U2F security key or hardware MFA device, the setup will be very similar. You get the code from the security key or the hardware device rather than from a mobile app. Next time, when you log in, you’ll use your regular credentials as well as an MFA password. Everything will continue to work just as before. It’s a simple and important step to safeguard your account.

Further, it is equally critical to implement these measures in your other platforms as well as in your personal life — after all, if hackers are able to gain access to your own digital resources, it could provide a gateway to exploiting those of your company. Startups are especially vulnerable here, since in the early days, personal and company assets often overlap and bleed into each other in both understandable and unexpected ways.

For startups, focusing on security can often be seen as a blocker. But setting up MFA is an essential step in your company’s growth — one that will safeguard your most valuable assets and set you up for continued success.

Xuan Gao is a Solutions Architect at AWS on the Startups team, where she helps startups grow their business and achieve their goals by using AWS efficiently. She is passionate about cloud technologies. She loves travelling and cooking in her free time.

How AWS Supported WOMBO’s Wildly Popular, AI-Powered App

AWS Editorial Team — Fri, 28 Jan 2022 16:42:14 +0000

When WOMBO, a Toronto-based synthetic media company, launched its AI-powered lip sync app at the end of February 2021, it took off like a rocket ship. In just ten months, 74 million people across more than 180 countries downloaded the app, making Wombo one of the fastest-growing consumer apps in history.

Today, through the WOMBO and WOMBO Dream apps the company has more than 10 million monthly active users, ranging from Gen Z teens through creators in their 90s—a point of pride for the company.

Ryan Khurana, Chief of Staff, WOMBO

“We get customer reviews from people saying, ‘My grandma doesn’t get a lot of entertainment on the internet, but she found your app, and she’s never laughed so much,’” says Ryan Khurana, WOMBO’s Chief of Staff. “That was a really important design principle. We wanted to make an app that was so simple that anyone could use it.”

For most startups, scaling quickly and becoming extremely popular is a desirable goal. But the rate of WOMBO’s growth exceeded even the founders’ wildest expectations, with 25 million downloads in the first month alone—something that created a huge challenge for the company.

“It just took off,” Khurana says. “Our AWS bill was higher, by an order of magnitude, than all the money we raised. If we did not have a partner that was as friendly as AWS, and as understanding of the situation, we would have gone bankrupt.”

Success didn’t have to come at a price though. AWS recognized the situation and stepped in to help the company lower its costs, manage its payments, and develop a better understanding of the AWS infrastructure. That allowed the company to continue to improve the quality of its algorithms and operate in the cloud, a central part of WOMBO’s vision.

“Our bet on the cloud is a critical part of what synthetic media is,” Khurana says. “If you compare how our app works to how the vast majority of competitor apps in this space, we’re the only ones that will bear the much higher costs of processing in the cloud, because we’re convinced that the real AI experience people want—these really unique and unexpected visuals—cannot be processed on someone’s phone in an appealing manner.”

On a technical level, AWS worked closely with WOMBO’s engineers to continually optimize the app’s data processing to make sure clusters didn’t get bottle-necked.

“The scalability of the AWS cloud in general was a real area of strength for WOMBO,” Khurana says. “To have someone like AWS understand the strategic partnership from the get go, help us out in the early days, and still continue to support us right now, our company wouldn’t be where it is otherwise.”

Now, WOMBO is entering a new phase and the company is expanding on their fun, easy-to-use AI. Its latest release in November 2021 is an AI art generator called WOMBO Dream which allows users to create unique designs that can be turned into digital and physical artifacts. The app has has received 10 million downloads since launch, with 6 million monthly active users this month alone, proving that WOMBO has tapped into a widespread demand for AI-generated content. Looking to what’s next, the company continues to invest in the identification and optimization of cutting-edge AI models for entertainment.

And while users currently share WOMBO’s content via social media platforms like TikTok, Instagram, and Twitter, the company also plans to develop its own content platform, which will serve as a new venue for creators and encourage further interaction with WOMBO’s products. It will be, Khurana says, “a new type of social media experience.”

That such ambitious goals are possible is a testament to how WOMBO has fundamentally transformed the way people think about AI. Before the company launched its lip sync app, many people viewed AI with a certain amount of trepidation.

“For millions of people,” Khurana says, “we’ve been able to give them this exposure to AI as something fun.”

Ultimately, Khurana says, that’s a valuable addition to the pace at which technological progress can occur. Machine learning for efficiency and productivity won’t be the only drivers of its broad-scale adoption, it also helps if it’s fun. And thanks to AWS, that’s what WOMBO has been able to provide.

Founder Security Fundamentals – Improved Security with Identity and Access Management

AWS Editorial Team — Wed, 26 Jan 2022 17:36:08 +0000

Authored By: Dave Barton, Principal Solutions Architect

At Amazon Web Services, we work with thousands of new startups every year, and we know how hard founders work to balance the demands on scarce time and resources. Often, a startup’s first goal is to build a minimum viable product, or MVP. This is a key milestone in their journey, where they can put their product into the hands of real users and start incorporating their feedback. Real users offer diverse experiences, reveal new use cases, and expose hidden bugs. And they also bring something very important: real user data. Users are trusting you with their information — which could be personal, financial, or medical in nature — and they expect you to treat it with care and respect.

When we speak to startups, we often talk about the concept of an MVSP, or minimum viable secure product. This doesn’t mean that a product must have every potential security risk mitigated at launch, just as it doesn’t need every feature implemented. By following AWS best practices, however, you can ensure that you’re building in a secure way and protecting your users and your business. While security is often seen as a blocker, putting the right guardrails in place can actually accelerate your development by allowing developers to deploy new features with confidence — reducing the amount of time they need to spend thinking about security and increasing the time they can dedicate to building features your customers will love.

Speaking of security guardrails, one of the first pieces of advice AWS gives to startups is, “Lock away your AWS account root user access keys.” In fact, that’s the very first item on our list of AWS Identity and Access Management security best practices — a sign of just how strongly we believe in this recommendation. Let’s start by understanding more about what the AWS account root user is and why you shouldn’t use it unless absolutely necessary.

So what is an AWS account root user?

From the official documentation, we can see that; “When you first create an Amazon Web Services (AWS) account, you begin with a single sign-in identity that has complete access to all AWS services and resources in the account. This identity is called the AWS account root user. You can sign in as the root user using the email address and password that you used to create the account.”

The root user is an integral component of an AWS account. There can only ever be one root user per account, and they have complete and unrestricted access to everything inside the account. The root user login is an email address, which must be unique across all AWS accounts in all regions. We strongly recommend that this email be an alias tied to a business domain, rather than a personal email address, and it should have a strong password and multi-factor authentication enabled. A handy tip here is to either use an email alias such as aws-production@yourdomain.com or, if your email provider supports it, you can use plus-addressing, which looks like aws+production@yourdomain.com.

There are two security fundamentals that highlight the differences between IAM users and root users in AWS:

Identity and Attribution – every action taken within a system should be attributable to a specific user.
Principle of Least Privilege – every user within a system should have the permissions to perform the tasks required of them and no more.

Looking at the AWS root account, you can clearly see that it fails both of these tests; if multiple individuals work using the root user, you cannot tell which individual performed which action, and the root user has full access to the account. This might feel fine when it’s just you building in your new account, but there’s a better way.

What is Amazon Identity and Access Management?

At its core, IAM is the AWS service that is used to manage access to all other AWS services. It is built around four foundational components:

Principal – the user or service (identity) that is being granted or denied access.
Action – the command(s) that are being allowed or denied. This could be very broad, such as full access to the Elastic Compute Cloud (Amazon EC2) service, or very tightly-scoped to a single command such as EC2:DescribeInstances.
Resource – the target to which access is being granted or denied. Again, this could be as broad as any EC2 instance or scoped to a specific instance.
Condition – an optional check that can be used to further limit permissions based on one or more criteria.

By combining these four elements, you can create an IAM policy that allows or denies access to AWS services using a consistent, logical, and readable syntax. If you want to learn more, we highly recommend watching this video by Becky Weiss, Senior Principal Engineer at AWS, where she dives deep on IAM and how to use it to protect your AWS environment effectively.

What is an IAM user, and why should you be using it instead?

Going back to the IAM documentation again, we can see that: “An AWS Identity and Access Management (IAM) user is an entity that you create in AWS to represent the person or application that uses it to interact with AWS. A user in AWS consists of a name and credentials.”

The key thing to note here is that an IAM user represents an individual or an identity. This allows actions taken by that user to be attributed to a specific person or application. Each IAM user can also have one or more IAM policies attached to them, which offers fine-grained control over the actions that user may take. In fact, those are the two key differences between the AWS account root user and an IAM user:

An IAM user represents an individual person or application, whereas there can only ever be a single root user for each account.
An IAM user only has the permissions granted to it by one or more IAM policies, whereas the root user has unlimited access, and there is no trivial mechanism for scoping down that access.

Let’s dive a bit deeper into the differences and similarities between the root user and an IAM user:

	AWS Account Root User	AWS IAM User
Has a password that can be used to access the console	Yes	Optional
Has an access key that can be used to programmatically access AWS	Optional	Optional
Can have a second-factor of authentication	Yes	Yes
Can be disabled	No	Yes
Actions taken by this user are attributable to an individual or application	No	Yes
Can be given limited access to resources in its AWS account	No	Yes
Can be given access to resources in a different AWS account	No	Yes

There are a few scenarios that make it clear why IAM users should be preferred, even in very small teams:

Using shared root credentials – if someone leaves your team, the credentials must be rotated, and everyone will need to be given the new credentials. If the credentials are also being used by applications, they will also need to be updated.
Root credentials leaked – while it’s possible to rotate the credentials of the root user, it’s disruptive (see above), and if the root credentials are leaked, the damage that can be done by an attacker is basically unlimited, including taking over or closing your account.
User error – when all users are working with the root credentials, the risk that they could perform a highly destructive operation by mistake is greatly increased. This could include deleting production services or data.

When and how should I use the root user?

We strongly recommend that you do not use the root user for your everyday tasks, even administrative ones. Instead, adhere to the best practice of using the root user only to create your first IAM user. Then securely lock away the root user credentials and add multi-factor authentication. This can be done in a few ways:

Hardware code generator or U2F token – this is a physical device that is required to gain access to the account to which it’s attached, and it is the most secure option for MFA. Once configured, the device can be stored in a safe or lockbox to offer additional security.
A software-based authenticator – this is an app, usually running on a mobile device, that generates secure one-time codes, which are used as part of the sign-in process. These are considered less secure than physical tokens but are still a good option for most use cases.
SMS-based MFA – this is considered insecure and is no longer supported for new users.

Some operations within your AWS account can only be performed by the root user. They are documented here, but the critical ones are:

With the above in mind, we recommend that you regularly (every month or so) check to make sure you can access the root user in your AWS account. Due to the high risk in using the root account, AWS Support has a high bar for evidence when it comes to recovering access. With this in mind, it’s a good idea to ensure that at least two people are able to access the root account, so that access is not lost if a single person is unavailable for any reason. You may also want to set up an alert so that you receive a notification when the root account is used, to ensure that it remains secure. Lastly, you should only use the root user via the AWS console, and you should not create access keys for the root user.

What’s next?

Now you have a solid understanding of what the AWS account root user is and why you should avoid using it for working in your AWS account. As a next step, go and see how the account root user is being used in your account today, and start working through our IAM security best practices. It might feel like a lot of work now, but doing it today while your team is small — and, ideally, before you have production workloads and real customer data in your account — can save you trouble, time, and money down the road. If you need help getting started, please reach out to your account manager or contact AWS Support.

Leveraging AWS Business Support as a Startup

AWS Editorial Team — Thu, 20 Jan 2022 14:35:25 +0000

Imagine this scenario: It’s the weekend, and your infrastructure team is making a key decision to scale your environment in order to support your customers and give them a top-notch experience with your product. At the same time, a newly-launched feature for your application is driving traffic to your website. Are you worried that it may cause downtime or slower response times while resizing parts of your tech stack?

You don’t have to go it alone. AWS Business Support can help you avoid problems and make sure your business is operating smoothly.

Whether you are testing out an integration with a new AWS service, requesting increases in service limits, obtaining best-practice guidance, or minimizing risk when making changes to enhance the customer experience, AWS Business Support can help you make time-sensitive decisions to keep your startup running smoothly. We offer a range of services, including troubleshooting errors, improving performance, providing AWS’s recommended best practices, and empowering your team to be successful while using our products and services.

How do I get in touch with Business Support?

Startup customers can interact with AWS Business Support via the support console. Each support case is classified according to severity, which ranges from General Guidance to Production System Down. The priority level you select should reflect the urgency of your request with respect to your infrastructure running on the cloud. Examples and a breakdown of severity can be found here.

Getting in touch with a cloud support engineer is the next step. AWS Business Support allows communication in three ways:

Call – Calling AWS Business Support puts you on the line with a cloud support engineer. Whenever you are having production issues or need to mitigate high-severity risk, we recommend using the call or chat option. If necessary, Business Support will use our screen-sharing tool to remotely view your screen to identify and troubleshoot problems.
Chat – Chatting live with a cloud support engineer is also a convenient way to address issues quickly, and makes it easy to share information such as resource IDs and error messages while having a conversation to get more insight into your support case.
Web (Email) – The web communication method allows you to have asynchronous email communication with a cloud support engineer. This is a useful way to ask questions about service features, obtain information about how to set up products, and receive general guidance.

Quick tip: If you would like to submit your support ticket in advance and have a conversation later, feel free to leave a phone number and a suitable time for our support engineers to call you. Email communications may cause time to be lost in critical situations. Therefore, it’s best to provide contact information whenever necessary.

Whats on the inside?

AWS Business Support teams are staffed 24/7 by highly trained engineers, and each support case is routed to an engineer who knows the ins and outs of the service in question. For example, a case revolving around AWS Lambda and Amazon API Gateway would be directed to someone who supports the serverless domain.

These are the same support engineers who work with our enterprise customers, so you can have confidence in the fact that as a startup you’re getting the same world-class level of assistance as our biggest customers. Support teams can also escalate requests to subject matter experts, and if needed, to the service team for product bug fixes and anomalies. These mechanisms help us provide the best in-house support while running your workloads on AWS.

What information do support engineers need?

Great support ticket writing helps ensure the fastest possible turnaround from the AWS Business Support team. A good support ticket includes relevant information about the issue at hand, which can help avoid some back and forth and allow the support engineers to get straight to solving your problem. Elements of a good support ticket include details such as the Amazon Resource Name (ARN), instance IDs, the timestamp of when you faced an issue, any relevant log files, specific Amazon CloudWatch metrics you may have observed, and error codes. These details give our support team a crucial head-start and help them identify exactly where the problem may lie.

But I’m just a small startup. Is it really worth the cost?

From a cost perspective, AWS Business Support starts at $100 per month. The price scales to 10% of monthly charges when AWS spend is up to $10,000, 7% for monthly AWS charges between $10,000 – $80,000, and so on. More details can be found on the pricing page. This asymmetrical cost relationship is advantageous for businesses, as customers are not charged on the number of tickets submitted or the number of conversations they have with AWS Business Support. We encourage customers to reach out with questions about any and all services that they use on our platform. Startups benefiting from AWS Activate also have credits toward Business Support which provide a great way to explore the offering.

Conclusion

AWS Business Support should be used to resolve blockers, answer questions, and help fix critical issues. The different severity levels and methods of communication allow us to tailor our response according to each scenario. Support cases are routed to domain-specific engineers, and we recommend providing detailed descriptions of problems for the fastest possible turnaround. These services are cost effective and can save startups hundreds of hours of time for a relatively small monthly fee. Think about it: How much does your engineering time cost?

AWS Business Support is here to help you build successfully, robustly, and according to best practices. We hope to interact with you soon through the AWS Support Center.

Parth Shah is a Startup Solutions Architect at Amazon Web Services. He enjoys working with startup customers in cloud adoption and business strategy as well as helping them design applications and services on AWS. Outside of work, he enjoys gaming, soccer, traveling, and spending time with his friends and family.

CelerisTx: Drug Discovery for Incurable Diseases with ML on AWS

AWS Editorial Team — Mon, 17 Jan 2022 15:11:17 +0000

Guest post by Noah Weber, CTO, CelerisTx, Christopher Trummer, CEO, CelerisTx, and Olajide Enigbokan, Startup Solutions Architect, AWS

Harnessing computational drug discovery on proximity-inducing compounds (PICs)

Humanity forged ahead with countless therapeutic solutions to treat disease. Recent technologies such as knockout (CRISPR – works at the DNA level) and knockdown (RNAi – works at the RNA level) have significant limitations. In contrast, Proximity-Inducing Compounds (PICs) combine the unique advantages of small molecules. Such beneficial features include oral bioavailability and ease of manufacture – along with those of silencing technologies such as CRISPR and RNAi. In contrast to RNAi and CRISPR, PICs affect proteins, thus providing a breakthrough therapeutic modality to target proteins, specifically associated with diseases. Alternatives that target proteins include biologics and inhibitors, and these alternatives have not expanded the range of pathogenic proteins that can be targeted. Over 80% of pathogenic proteins are associated with a disease not yet amenable to pharmaceutical interventions.

Targeted protein degradation (TPD) is the most popular proximity-inducing drug modality, first demonstrated in the early 2000s. It also received a Nobel Prize in Chemistry for discovering ubiquitin-mediated protein degradation in 2004. The TPD concept involves the selective degradation of proteins by hijacking internal cellular machinery within the human body. Compared to occupancy-driven methods such as inhibitors, TPD mechanisms have the advantage of eliminating scaffolding functions and thus addressing the root cause of the problem rather than just treating symptoms. In general, tools to eliminate misfolded and denatured proteins within a cell are needed to maintain biological homeostasis.

While various machineries can be actively utilized via PICs, such as lysosomes and autophagy, Celeris Therapeutics focuses on the ubiquitin-proteasome system. Based on this system, unwanted proteins are tagged with ubiquitin, a small signaling protein. As a result, this process initiates the degradation of pathogenic proteins.

However, the assessment for TPD in an experimental laboratory setup is slow and expensive. Celeris Therapeutics, therefore, implemented a computational workflow to predict protein degradation effectively and thus shorten and streamline the drug development timeline.

Proximity-inducing compounds

Celeris Therapeutics (CelerisTx) is pioneering the adoption of AI on PICs, focusing on Targeted Protein Degradation. In terms of data, in protein degradation, that means identifying multiple molecules that interact with each other in a particular way. Imagine LEGO® bricks. Instead of fitting them together according to only a single dimension i.e., shape, these elements have multiple dimensions or requirements to satisfy. Some are chemical, biological, physical-driven, etc., meaning, that the interactions are subject to specific laws. These laws need to be quantified and embedded in data. Concrete examples of such information include 3D structures of molecules represented as graphs, hydrophobicity, or electrostatic potential. Altogether, this amounts to roughly 20 different features that experts in TPD curate. As a result, an enormous amount of data is being generated that needs to be analyzed using machine learning algorithms and, specifically, geometric deep learning pipelines. A concrete ML application is required to determine the degree of interaction between molecules involved in a so-called ternary complex. In other words, considering two proteins, we expect a scalar value that determines the degree of interaction. Beyond being accurate and generalizable, we have to make rapid predictions in a performant system. As we simulate different protein-protein interactions, it is necessary to discard inaccurate predictions swiftly to parse through the space promptly. This has proven to be a challenge for an interaction dataset with more than 20,000 protein pairs.

How AWS was leveraged for various PIC discovery projects

To save developer time and gain quicker insight into the particular problems surrounding machine learning development, we turned to Amazon SageMaker. Its capabilities allowed us to avoid implementing some machine learning infrastructure solutions ourselves, such as bias detection or hyperparameter tuning. Bias detection was relevant during data preparation; hence, we leveraged SageMaker Clarify. This was instrumental in ensuring the quality of our data, before we even started modeling. Afterward, once we had begun modeling, it was important to have direct integrations of geometric deep learning libraries, such as SageMaker Deep Graph Library (DGL). Since DGL is an open-source python package for deep learning on graphs, we used it to gain a quick setup of the infrastructure needed for geometric deep learning, which is essential.

Once the initial models were set, to find the best hyperparameters quickly, we used SageMaker Automatic Model Tuning. We estimate that we thereby avoided months of development time in coding up the hyperparameter optimization frameworks. We leveraged SageMaker Experiments to track and organize all changes to the experiments. This solution was especially important. Consider the aforementioned problem of determining the strength of the interaction between proteins involved in the ternary complex, the 15+ geometric deep learning architectures that we have experimented with from attention layers, the pure convolution operators on graphs, etc. The trackability of the different models and their respective parameters was important, to determine what would eventually work. Once we had created the final version of the model, we needed to debug them and further optimize the pipeline. For this, SageMaker Debugger proved helpful.

Along with software solutions from SageMaker, we needed a great deal of computing power for our vast data and deep learning pipelines. Hence, we built an optimized Spot Fleet solution for the GPUs used in training our machine learning workflows.

Optimizing the cost of our ML pipeline

Spot Instances are a special kind of Amazon EC2 instance that give 90% discount compared to On-Demand prices. Sometimes, due to high demand for computing resources, computing resources can be interrupted, causing a user to lose a whole compute session. It is crucial to save the computation’s intermediate results, if one trains a deep learning model for a long time, or if some necessarily long-running computation was executed on the Spot Instance. Aside from saving the computation’s intermediate results, the user should also be able to automatically continue the computation on another Spot Instance from the same intermediate step where the disconnect happened.

The AWS SDK for Python (boto3) was used to create a script that will, upon execution, send a Spot Fleet request to create a GPU Spot Instance with Amazon Machine Images (AMI), containing all of the software and dependencies to execute our machine learning code. Further, we define user data scripts that download a docker image from Amazon Elastic Container Registery (Amazon ECR), and then Git clones the code from our repository. Finally, the user data scripts executes the geometric deep learning pipeline code. This code has been modified such that after each epoch of machine learning training, it sends datasets, logs, models, and checkpoints to an Amazon EBS volume. To ensure that the objects are saved, we dump them all to Amazon S3. Aside from defining a Spot Fleet, we have also enabled replenishing any Spot Instances that disconnect in the process, which allows us to continue the training automatically from the last checkpoints found in the S3 buckets.

SageMaker offers something very similar, called Managed Spot Training. However, we did not leverage the SageMaker offering, because, when used as part of SageMaker, the EC2 instances are more expensive than standard ones. This makes sense, since SageMaker provides a lot of ML offerings, ease of use, less engineering time needed, etc. For us, architecting this solution was a strategic investment of time, because we knew we were going to use GPUs in different capacities for multiple years.

HPC was another significant use case that we needed to fulfill with AWS. The huge data space that needs to be parsed with geometric deep learning comes with a substantial computational burden. We scaled up our computations horizontally with AWS ParallelCluster and created a shared filed system, Amazon FSx for Lustre, so that all compute nodes could access and modify data from the same place. Also, AWS supports Slurm as a workload manager, which allows us to distribute the work across CPUs and GPUs at the same time. This eventually allows not only faster experimentation but also optimized model serving. Again, one should note that SageMaker offers a similar service for large datasets and models, called Distributed Training Libraries.

The future of proximity-inducing compounds

Preliminary results from our pipeline are available for Parkinson’s disease progressing to preclinical studies. Experimenting quickly and reliably is essential for computational drug discovery as we continue to improve our drug discovery pipeline with new architectures and approaches. Our R&D approach, in which we invest substantial resources in research, means that we face many ongoing risks we need to mitigate. AWS standard offerings are one way to ensure we are agile and move faster from research to lab and then to market.

Conclusion

Now is the time to increase productivity in drug discovery. The inverse of Moore’s Law in pharmacology, known as “Eroom’s Law,” indicates research and development costs required to develop a new drug will only continue to rise. An approach to address this is to harness and streamline the process of performing sustained machine learning experiments for drug discovery.

	Christopher Trummer is a Co-Founder of Celeris Therapeutics and serves as CEO. He has been an invited keynote speaker in AI for drug discovery conferences multiple times and is co-author of peer-reviewed publications in various journals.
	Noah Weber serves as Chief Technology Officer at Celeris Therapeutics. He is a Kaggle Grandmaster and adjunct lecturer at the Vienna University of Technology and the Vienna University of Applied Sciences.
	Olajide Enigbokan is a Startup Solutions Architect at Amazon Web Services. He loves working with startups (most especially builders) to discover the value of the AWS cloud

What Amazon CTO Werner Vogels’ Predictions for 2022 Mean for Startups

AWS Editorial Team — Wed, 12 Jan 2022 16:32:47 +0000

As Amazon’s CTO since 2005, Werner Vogels has observed macro trends around the technology industry, giving him a unique perspective to distinguish substantive progress from mere fads. He recently published his views on what he sees in store in 2022 for cloud technology and the technology world in general.

In his post, Werner lays out five core predictions, regarding the growth of artificial intelligence (AI), the abundance of data, the power of machine learning (ML), architecting for sustainability, and the full reach of connectivity via the Internet, backed by cloud-based resources. I think his analysis provides takeaways for startups overall and where they might seek to create value in the next year.

Let’s dive into each of these anticipated developments and their potential impact on the world of startups.

1. AI-supported software development takes hold

In this prediction, Werner highlights exciting AI advancements that are just starting to impact the world of developing and running software at scale. We see this in AWS tools such as Amazon DevOps Guru, which brings to bear ML capabilities for observability and operational best practices, and Amazon CodeGuru, which uses ML to inspect a developer’s code for issues. We also see this in the capabilities of AWS partners’ tools, such as GitHub Copilot, which uses ML to assist developers in writing code.

A developer might soon leverage ML and AI capabilities to extend far beyond their own immediate experience and knowledge, magnifying their impact and ability to deliver new features and capabilities to their core business products.

For startups, the impact of this could be massive. Imagine accelerating application development via code generation tools, inspecting the applications via tools that alert you to potential bugs, and then deploying and managing the applications with automated guidance on performance or cost improvements. Time to market and iteration cycles would shrink drastically. Production issues could also be corrected more quickly. Performance could be tuned before customers are impacted. The result would be even leaner startups, able to achieve greater development speed and focus, a key to success and growth.

2. The everywhere cloud has an edge

As of 2021, almost 60% of the world’s entire population has access to the internet. Internet connectivity is pervasive inside our homes, the businesses we frequent, and even our modes of transportation. Increasingly, the devices that we carry, wear, or have around us at any time could also be connected to the internet through the cloud. To Werner, what this means is that opportunity presented by the scale and size of the cloud to allow companies to build software and devices that live alongside us is greater than ever.

The cloud now exists not just in places like AWS’s geographic regions and edge locations, but in all things connected to the internet. Technologies such as those in the AWS IoT family, and AWS products that help make the management of distributed networks of devices easier to operate and scale, give developers new superpowers in this “everywhere cloud” world.

This creates an enormous opportunity for startups to build advanced technical products and services that impact the daily lives of people around the world. From mom-and-pop, brick-and-mortar small businesses, through global infrastructure and logistics networks, to home devices or wearables, the cloud at the edge concept will support many new product ideas. Think advanced AI/ML capabilities at the edge guiding business decisions from afar, or sensor-driven devices that process anomalies in near-real time, potentially reducing costly downtimes and maintaining production of needed goods.

3. The rise of smart spaces, especially in senior care

Around the world, the average age of individuals is increasing. Advancements in medical practices and living standards globally over the last century have increased life expectancy, compared to earlier times. Combined with changes in family planning in many countries, this means fewer younger people, as well. This trend leaves us with a gap in elder care that presents an opportunity for new technologies to aid not just the elderly but those who support and care for them, through the rise of smart spaces.

Taken with the growth of the everywhere cloud, the potential for AI and ML to impact daily life creates whole new opportunities for startups. From developing smart wearables and household devices that work differently for older family members to implementing technologies that help them access and more easily engage with the world around them, companies can change how we age with dignity and independence, and how we support the elderly.

4. Sustainability gets its own architecture

Developers building applications in the cloud are familiar with the concepts of tuning for performance metrics such as page load time, API latency, or the total cost of infrastructure. Tools, practices, and guidance exist to help them make the best decisions regarding each of these concerns. But one common area has been less emphasized in the past—the impact of a workload with regard to sustainability. How does scaling to handle specific workloads impact our environment?

In this prediction, Werner describes how developers will increasingly consider the environmental impact of their applications and how to potentially tune for it, alongside performance and cost. With the recently announced AWS Well Architected Framework pillar on Sustainability, and the data that will be made available by the upcoming AWS Customer Carbon Footprint Tool, our customers can now better understand and architect for this.

This means startups can now assess the environmental impact of their business. For some, this will become a championing point, used to position the business toward new customers and for hiring. For others, it will represent a tracked business metric that drives innovation and product capabilities. One can imagine venture capital groups that focus on environmentally conscientious companies. There will also be opportunities for new products that consume this carbon footprint data, as well as that made available via programs such as the Amazon Sustainability Data Initiative, that can lead to impacting the lives of end consumers.

5. A new wave of connectivity initiates a new class of applications

As I wrote earlier, about 60% of the world’s population has access to the internet. While this number represents the fastest-growing technological advancement to humanity in history, it still leaves behind a huge number of people. The forecasted rapid expansion of low Earth orbit satellites that can deliver high-speed connectivity to nearly every point on the planet will help close the coverage gap, making the internet truly global.

In a sense, this unlocks even more opportunity for the predictions covered previously. When the internet is literally everywhere, and the cloud extends to all of its edges—with access to AI and ML tools, data analytics and processing capabilities at massive scale, and tools to enable smart devices anywhere—whole new classes of business opportunity will emerge. For startups, this means new addressable markets and new ways to challenge existing norms and push the bounds of what exists today.

This also opens the door for the growth of startups in these emerging markets. We’re already seeing investment increase across major parts of the world, for example the growth of Y Combinator cohorts in Africa and Southeast Asia. As a result, developers in these newly connected parts of the planet will be able to impact the world directly around them. We are likely to see cultural and societal impacts, as well; increasing the ability to communicate and share information makes humanity ever more connected.

Looking ahead

These trends suggest incredible opportunities for startups in the new year and beyond. The maturity of the ecosystems around some of these technology advances is approaching a threshold of common place that opens doors for new companies to have major impacts and success, quickly and with minimal investment. For some companies, the near-term trends will spark transformative business ideas that could impact global communities. Developers have more power at their fingertips to build products quickly and at an unprecedented scale. By leveraging the capabilities of AI and ML, processing data on a massive scale, expanding your product’s reach to the furthest extents of the internet, and building truly human-friendly interfaces, you can change the world.

As always at AWS, we’re energized to be here with you at this exciting time and look forward to helping you build.

Chris Munns is the Tech Lead & Advisor for the Startup Solution Architecture organization at Amazon Web Services. Chris works with peers at AWS on how to better support AWS’s startup customers and directly engages with helping hot startups overcome complex technical challenges. In a little over 9 years at AWS, Chris led Developer Advocacy for AWS Serverless technologies, was the global Business Development Manager for DevOps technologies, and was a Solutions Architect in the early days of the AWS field. Before AWS, Chris held senior operations engineering posts at Etsy, Meetup, and other NYC based startups. Chris has a Bachelor of Science in Applied Networking and System Administration from the Rochester Institute of Technology

Why Every Startup Should Set Up a Budget — and How AWS Budgets Makes It Easy.

AWS Editorial Team — Mon, 10 Jan 2022 14:03:25 +0000

Do I really need a budget?

But in a space where one surprise bill can break the business, freshly launched startups must be frugal from the get-go and keep a close eye on costs in order to maximize runway. According to CBInsights, the number one reason startups fail is because they run out of money – so not establishing good cost-control habits from the beginning could end up being a big mistake.

Although monitoring costs on AWS might seem like an arduous task (and make it tempting to ignore the process altogether), it doesn’t have to be. With AWS Budgets, it takes just a few minutes to set up a budget, which can help you catch surprise bills before they happen. You’ll also be able to monitor costs and usage over time, allowing you to optimize your monthly bill and maximize usage of the perpetual AWS free tier once you’ve transitioned off of credits.

What is AWS Budgets?

AWS Budgets lets you set budgets to track your costs and usage, and you’ll get notifications when budgets are exceeded at self-defined thresholds. AWS is built to handle workloads of any size, from early-stage startups to full-blown enterprises, so it’s important to have a method in place to catch cost overruns or unexpected spend before you reach enterprise size. Putting a budget in place early on sets you up for success and helps you build good cost-hygiene habits from the start.

Set it up

1-click deployment

Setting up a budget through the AWS console will take no more than a few minutes. If you’d like to follow along and learn the process, start at step 1. Otherwise, you can 1-click deploy a budget with the AWS CloudFormation template below. If you haven’t used CloudFormation templates before, follow along here.

Region	View	View in Designer	Launch
US East (N. Virginia)	View	View in Designer
US East (Ohio)	View	View in Designer
US West (N. California)	View	View in Designer
US West (Oregon)	View	View in Designer

Step 1 – Select a target amount

The first step is to establish a monthly budget value. Here’s how to determine what that value should be:

Do you already have a set amount to spend monthly on infrastructure? If so, use this for your budget value.
If not, have you run workloads on AWS for the last six months? If so, go to the AWS Cost Explorer home page, and the default view (shown below) will show your overall spend from the last six months. Take the average of this monthly value, and use this figure for your budget value.

If you have less than six months of spend history, figure out your pain point. What amount of monthly spend would break the bank? Use this figure for your budget value.

The purpose of a budget is to alert you before an unexpectedly costly bill causes substantial financial hardship. But once you start receiving notifications that you’re exceeding your targets, it means you’re already headed down a perilous path. For a budget to be effective, you’ll want to come up with a sweet-spot value where you’re not getting too many notifications, but just enough to provide visibility and give you time to correct any accidental misconfigurations.

Step 2 – Create the budget in AWS Budgets

After you’ve determined your target value, go to AWS Budgets and create a budget. Since the goal is to make sure you don’t spend too much money across all services, choose the Cost budget, which will keep track of your dollar spend.

Select the following options when setting the budget amount:

Period: Monthly
Budget effective date: Recurring budget
Choose how to budget: Fixed
Start month: Leave as default
Budgeted amount: The target amount you determined above

After setting the budget amount, you’ll have the option to scope your budget, which lets you create budgets that apply only to specific sets of services, tags, regions, accounts, and more. For this example, however, the budget you’re creating is acting as your “runaway spending budget,” which is intended to catch cost overruns at a high level. Therefore, no scopes need to be applied.

However, as a general best practice you should have a separate budget for every workload you’re running. You can use the budget scopes to help monitor things like testing and development costs on new services or Amazon Elastic Compute Cloud (Amazon EC2) instance types, or to set budgets for different teams within your organization. Additional examples of custom budgets you can set to avoid unexpected charges can be found here.

Step 3 – Set up notifications

You can configure alerts so that you’ll get notifications if your spending reaches defined thresholds. You want to monitor spend, but you also don’t want so many notifications that you start ignoring them. Therefore, we recommend that you start by creating two alert thresholds, so that you’ll receive an email when you’ve hit 50% and 75% of your budgeted amount.

For a more advanced setup, you can pipe notifications across other channels via Amazon Simple Notification Service (Amazon SNS), which will lets you expand notification methods to SMS, HTTP endpoints, and more. If you like to use Slack, you can connect Amazon SNS to AWS Chatbot to get your budget alerts sent directly to your Slack channel. For details on how to connect your Amazon SNS topic to AWS Chatbot for Slack, follow along here.

Conclusion

Regardless of size or funding stage, every startup should have a budget to catch runaway spend at an account level. By investing just a few minutes of your time to set up this budget, you’ll be able to avoid cost overruns, maximize your runway, and get in the habit of monitoring costs over the long term.

And while establishing this budget is an important step, it’s not the only action you can take toward managing costs. Once you’ve established your first budget, consider expanding to include other budgets that track different dimensions – like tracking specific costs to each developer, or making sure you stay under a certain amount of machine learning training hours on a particular Amazon EC2 instance type. Even as your startup grows in size and complexity, AWS Budgets makes it fast and easy to keep tabs on costs without needing to invest hours tracking them by hand – and can help you avoid potentially expensive mistakes in the future.

Melissa Kwok is a Solutions Architect at AWS, where she helps customers of all sizes and verticals build cloud solutions according to best practices. When she’s not at her desk you can find her in the kitchen experimenting with new recipes or reading a cookbook.

The Impact of Compliance on the SaaS Sales Cycle

AWS Editorial Team — Thu, 06 Jan 2022 18:48:54 +0000

By Jon Topper, Founder and CEO, The Scale Factory (an AWS SaaS Competency Partner)

Selling to enterprise businesses can involve months-long sales cycles and detailed procurement questionnaires that take up a CTO’s time and attention. AWS can help make the process more efficient.

The journey to Series A and beyond

In the early days of a B2B Software as a Service (SaaS) startup, everything is about reaching product-market fit, to achieve annual recurring revenues (ARR) of $1 million as efficiently as possible. Because this means iterating quickly on product features, nonfunctional requirements, such as security and operations, often take a back seat.

As a startup journeys to an ARR of $1 million, venture funds may consider making a Series A investment. Along with money, such an investment comes with targets: at this stage, the company needs to aim for ARR of $5 million or $10 million.

With these bigger ARR targets, involving the founders in every sale makes little sense. Also, by now, the product and commercial model are usually pretty well understood, and so dedicated sales hires can take them to market.

Enterprise buyers are valuable but different

Enterprises spend more money than small and medium-sized businesses (SMB) do on their SaaS platforms, especially where those are priced according to usage. Research by Redpoint shows that annual customer churn is just 6%–10% for enterprise buyers, compared with 31%–58% for the SMB segment. Enterprises therefore spend more and are more likely to remain customers longer. Good SaaS salespeople will prioritize such sales.

But selling to enterprises is very different from selling to SMB or mid-market customers. In a small business, the purchasing process often involves one decision maker, with access to their own budget. An enterprise, however, has annual and quarterly budget cycles, all managed by a procurement team, with representatives from the finance, legal, compliance, and technology departments.

While the buyer-stakeholder you’ve been working with has a business context for the purchase of your product, this might not be true of a procurement team. Rather than having a nuanced conversation, for example, their primary mechanism for engaging with you will be a supplier questionnaire. Sometimes called an RFP or RFI document, this is often extensive, with hundreds of questions about topics ranging from how your out-of-hours support works to what kind of locks are on your office doors.

Even if your service won’t hold sensitive personal or commercial data, you’ll need to be convincing in your answers to these questions in order to make the sale, because it’s easier for a procurement team to apply the same standards to absolutely everything they buy than to have a deep contextual understanding of each purchase.

The first time you’re presented with a procurement questionnaire, it usually lands on the CTO’s desk, taking up days or weeks of time they could be spending on more strategic activities.

Getting compliant helps drive sales success

Procurement questionnaires exist for one reason: to reduce risk to the enterprise. One way to help reassure enterprise buyers that you care about managing risk is to adopt an industry standard, such as ISO 27001 or SOC 2.

ISO/IEC 27001 is an internationally recognized standard that establishes requirements for an information security management system. Meeting the requirements of this standard can help organizations keep financial, intellectual property, and employee information secure.

SOC 2 is the System and Organization Controls (SOC) for service organizations, which includes a set of audit reports providing evidence that an organization conforms to the standards they set for themselves around a set of defined criteria.

These two standards aren’t mutually exclusive: you can use SOC 2 to report on compliance with standards set using ISO 27001.

Adopting either standard requires an organization to look critically at how they manage information security, to improve this management, and to document their processes.

Teams who have invested in this level of compliance tend to find that procurement questionnaires become easier to fill out, because the team has already thought about the issues of concern to their buyers. In some cases, having ISO 27001 or SOC 2 documents to share with a customer can reduce the amount of the questionnaire to be completed, or even remove the requirement entirely. In some cases, a buyer will require one of these compliance levels before they’ll even speak with you.

FundApps is a Fintech company that engaged The Scale Factory to help design a new AWS tenancy model for their B2B SaaS platform. They completed an SOC 2 audit in 2020. CTO Toby O’Rourke mentioned that once they had done so, complex questionnaires and follow-up negotiations that regularly took weeks were frequently completed in half a day, shortening the sales cycle by as much as a month. Further, he’s been able to delegate the procurement paperwork to the head of information security, freeing up his own time for more strategic activities. I’ve heard similar stories from the leadership of other SaaS companies with which we’ve worked.

So, our advice is to invest in becoming compliant, in order to drive greater sales success. And you can do it on AWS.

AWS can help with compliance

The AWS cloud provides the pieces to build a secure, compliant platform that keeps your customers’ data safe. In the shared responsibility model, AWS takes responsibility for the security of the cloud, while you take responsibility for security in the cloud. You can achieve that by designing your architecture according to the best practices documented by AWS Well-Architected.

A Well-Architected review from an experienced AWS Consulting Partner can help identify ways to improve the security of an existing platform. There’s even a dedicated SaaS Lens, which considers issues unique to multi-tenant environments. When The Scale Factory engaged with FundApps, we used a Well-Architected review as a starting point.

AWS Control Tower

Companies with multiple AWS accounts, or even those just starting out with AWS, can use AWS Control Tower to set up and govern their AWS environment. AWS Control Tower establishes a centralized landing zone, providing identity protection, security threat detection and alerts, log aggregation, backup, and other functions. This centralized management and governance can help lay a secure foundation for your cloud resources.

You can set up the landing zone with best-practice blueprints for security and governance. When you vend a new AWS account from Control Tower, these blueprints are used to apply configurations. The standards in the blueprints can be created by your security team or with the help of a Consulting Partner, such as The Scale Factory.

Good security relies on strong identity and access management. AWS Control Tower can automatically configure AWS Single Sign-On in your new accounts, setting appropriate access controls for your teams, linked to your corporate directory, if appropriate.

To prevent, or be alerted to, particular types of misconfigurations, Service Control Policies, organizations policies that limit the permissions users can be granted, can be applied to your accounts. These guardrails stop those configurations from being made at all, even by root users. Alternatively, AWS Config, a service for managing configurations of your AWS resources, can monitor for configuration changes and either send alerts about noncompliant events, or run automation to take corrective action. All of these settings can be rolled out automatically to all new accounts in your landing zone.

Some compliance regimes require that you record and store system log entries for long periods of time. In this case, landing zone blueprints can set up AWS CloudTrail logs to be delivered to a separate security account, which only your security team can access to read, but which no one can modify. Other centralized security detective controls can also be set up, to monitor all of your accounts from a single interface.

Blueprints are not limited to security and compliance. They can be used to set up anything you like in your new AWS accounts by default. For example, perhaps you want to use HashiCorp Terraform Cloud to provision your applications. Control Tower can be configured to set up that integration for each new account. This could be a time-saver, if your SaaS tenancy model involves creating a fresh AWS account for each customer.

Learnerbly is a fast-growing workplace learning platform that engaged The Scale Factory to help them move closer to compliance with ISO 27001 and PCI DSS. We built out a secure landing zone for them, based on our B2B SaaS Foundations product from the AWS marketplace. Already a few steps closer to their compliance goals, Learnerbly can now also onboard new developers in seconds, a necessity for future growth.

Conclusion

To increase the chance of success when selling SaaS to enterprise customers, consider working toward compliance with ISO 27001 and/or SOC 2. Deploying AWS Control Tower provides a lot of the groundwork required for controls under these regimes. While it won’t get you all the way to compliance on its own, it does streamline and facilitate the process. Also, working with AWS SaaS Factory and/or an AWS SaaS Competency partner can help you reach more reliable outcomes, sooner.

Jon Topper is founder and CEO at The Scale Factory, where he’s been helping teams deliver business value from their infrastructure platforms since 2009. The Scale Factory is the only AWS Consulting Partner in the world focused exclusively on helping SaaS companies achieve hypergrowth. As well as his day job, Jon co-organizes, and hosts the AWS User Group UK. For this and his other community contributions, he’s been recognized as an AWS Partner Ambassador.

About AWS SaaS Factory

AWS SaaS Factory provides business and technical advisory to organizations at any stage of the software-as-a-service (SaaS) journey. Whether looking to build new products, migrate existing applications, or optimize SaaS solutions on AWS, the AWS SaaS Factory Program can help. Please reach out to your AWS account representative to inquire about engaging AWS SaaS Factory.

How AWS Helped PostEra Scale a Global COVID Antiviral Project

AWS Editorial Team — Fri, 17 Dec 2021 15:36:49 +0000

Guest post from Matt Robinson, CTO of PostEra: A company applying its latest advances in machine learning to drug discovery.

Early in the pandemic, while the world was largely focusing on other important technologies and therapies to combat COVID, those of us in antiviral development felt that the development of a pill to treat COVID was a necessary plan to pursue. With concerns ranging from public anti-vaccination sentiments to the difficulty in delivering vaccines or antibodies to the Global South, we saw a need for an accessible therapeutic treatment. But we had trouble getting anyone to listen.

Recognizing this, in March 2020, we at PostEra helped launch a worldwide effort to develop novel antivirals, using technologies from AWS, one of the early supporters of this ambitious project. Governmental organizations were slow to fund it, directing their resources elsewhere. AWS staff, however, rolled up their sleeves and helped us rapidly scale the infrastructure and innovative technologies to speed compounds toward clinical development.

Map of the global contributors to the Moonshot consortium.

The goal was a therapeutic pill that would be cheap, easy to stockpile, deploy, and administer. At times, it seemed like a moonshot.

In fact, that’s its name: the COVID Moonshot, and it has since become the world’s largest open-source drug discovery effort. This initiative was unique by borrowing an open source model from software development. COVID Moonshot accepted input from labs in the UK, Israel, US, China, India, and Ukraine. It unleashed the power of machine learning to speed up the process of molecular design by assessing thousands of crowdsourced ideas from chemists to determine which compounds were the most easily makeable. While chemists were stuck at home around the world, machine learning was generating predictions and simulations that compressed weeks of human endeavor into a matter of hours, prioritizing what needed to be made

Building the infrastructure for a global collaboration

PostEra uses our latest advances in machine learning to improve the way we discover new pharmaceutical drugs. From a drug discovery perspective, the COVID Moonshot required a new type of architecture, to combine the ideas of a completely distributed team amid global lockdown with these latest machine learning advances.

In AWS, we were able to change our infrastructure quickly to address these challenges, with services scaling to analyze the rapidly arriving data. Solutions architects provided additional insights on building an architecture where reacting quickly to project needs would be straightforward. About a week after the decision was made to launch the initiative, we had a new website up and running to serve the hundreds of chemists contributing designs.

The initial website provides a way for chemists around the world to input ideas to the platform. A graph shows the number of submissions over time.

Machine learning helped scientists predict which molecules would be easy to order or make in a lab. Using Amazon Relational Database Service (Amazon RDS), we were able to quickly spin up a multi-terabyte database of more than 10 billion catalog compounds to search. Additionally Amazon GPU G-class spot instances, which are readily available, as well as services such as Amazon Elastic Container Service (Amazon ECS) and Amazon Elastic Kubernetes Service (Amazon EKS), proved ideal for building the machine learning infrastructure needed for this technology to scale and provide the necessary predictions..

The initial fragment screen performed by collaborators at Diamond Light Source identified promising compounds that nicely fit in the protein.

Eventually, these tools were developed into the PostEra Manifold platform. Manifold runs completely on AWS, with an isolated, reproducible, and customizable architecture. The COVID Moonshot initiative has gone on to receive a more than $10 million grant from the Wellcome Trust to continue developing anti-SARS-CoV-2 therapeutics, work facilitated by how quickly we can iterate on our computational tools with AWS.

Since Moonshot launched, we’ve continued to innovate using the AWS infrastructure stack. Moving our tools over to AWS lambda allows us to handle workloads that seemed almost unimaginable a couple of years ago. Our clients can assess the synthetic accessibility of more than 100,000 molecules in an hour. We continue to integrate further AWS tools, to ensure the utmost security and utility for our clients, who can integrate their own custom data. The infrastructure that powered a distributed global initiative of antiviral drug discovery scientists is now helping to drive biotech and pharma drug discovery projects more broadly.

Meanwhile, the need for antivirals seems more urgent than ever. Scientists warn that COVID is becoming endemic. That means, we’re going to have to live with it. And that won’t be easy, even with the development of lifesaving vaccines and monoclonal antibodies. Storage and delivery difficulties, access issues, and anti-vaccine sentiment continue to impede global vaccination efforts. All the while, the Delta variant and others keep raging through unvaccinated groups, even breaking through vaccination.

Recent results from trials indicate that antiviral pills could provide crucial treatment, while adding a needed tool for variants that evade vaccines and antibodies. COVID Moonshot is still hard at work in continuing this mission, as we advance these molecules further towards human trials. Meanwhile, we at PostEra, using the resources of AWS, continue to refine machine learning technologies, such that the drugs of the future can get to humans more quickly.

Matt Robinson is the Chief Technology Officer at PostEra. As both a software developer and drug discovery scientist, he works to bring the latest advances in machine learning and computing to the biotech industry. He previously did research in biomolecular simulation and machine-learning for drug discovery at Yale University and University of Cambridge, before co-founding PostEra with his graduate school advisor.

ICONY: Detecting and Handling Fake Accounts with Amazon Fraud Detector

AWS Editorial Team — Wed, 15 Dec 2021 21:47:30 +0000

Guest post by Uwe Thomas, Managing Director at ICONY GmbH, and Anna Grüebler, Senior AI Specialist Solutions Architect at AWS.

In our digital world, as the popularity of online dating platforms and websites continues to grow, so too does the threat from scammers, bots, and other bad actors. ICONY GmbH, a white-label dating platform based in Germany, helps address this issue by rigorously validating users — allowing its business partners to launch their services with a database of reputable and up-to-date profiles already in place.

Today, more than 200 media companies partner with ICONY to create dating platforms under their own brands and/or domains. For each partner, ICONY sets up the look and feel of the platform and also manages its day-to-day operation — handling everything from additional application development, to editorial and customer support, to monthly billing of premium memberships and issuing credits to partners. ICONY also provides partners with a transparent backend showing daily behavior, user conversions, and sales generated on the platform.

Having a community of real, active users — not bots or fake accounts — is critical to the success of any dating platform, and so far more than 1.5 million people have registered with ICONY and its partners. Maintaining the trust of these individuals is of the utmost importance, which is why ICONY continually invests in new methods to detect and remove fraudulent profiles from the network. This leads to a better user experience, fewer abuse reports, and higher retention.

Initially, like many other online platforms, the ICONY support team relied on abuse reports to identify fake users and scammers, and then added internal rules and had members of the team review sign-ups. Recognizing and adjusting to the changing strategies by bad actors was a constant struggle, however. Each time a new type of fraudulent behavior appeared, the support team needed to manually create a new rule to address it — a very time-consuming endeavour that could not be scaled up and sustained. To solve this problem, ICONY decided to take a more sophisticated approach to fraud detection, using machine learning and the Amazon Fraud Detector service.

Amazon Fraud Detector provides everything that’s needed to build, deploy, and manage fraud detection models. To train the model, they used ICONY’s historical data of legitimate accounts, as well as the fraudulent accounts the support team had previously identified. Together with other AWS services, ICONY was able to create a bespoke fraud detection solution that did not require in-house machine learning expertise. After a few days of planning and discussion, the coding and integration of fraud detection on the platform was completed in just two days.

The following two diagrams show the high-level overview of the solution:

The implementation is divided into two steps: training and deploying the model, and evaluating and handling new users. The steps are as follows:

Labeled historical data of examples of fraudulent and legitimate transactions is loaded from Amazon Simple Storage Service (Amazon S3).
Using the historical data, a model is trained in Amazon Fraud Detector. This model will be used to evaluate new transactions and returns a risk score.
The model offline metrics can be reviewed by ICONY’s support team after training. These metrics show for each risk score the associated true and false positive/negative rates based on the training data. Risk scores are in the range of 0–1,000, with 0 representing the lowest possible risk, and 1,000 indicating the highest possible risk.
ICONY’s support team define the thresholds of the risk score returned by the model for different outcomes. The support team selects appropriate risk score cut-off values based on their business knowledge. In this case they defined three outcomes: “allow”, “monitor”, and “block” and assigned risk score cut-off values for each. These are the conditions on how to interpret the risk score during fraud prediction. For example, a risk score greater than a certain value can be defined to return the outcome “block” This automates the behavior of the web front-end.
The model output and business rules are defined in a Detector. The Detector makes the decision on what the outcome is. The Detector is deployed as a managed API endpoint that scales automatically with demand and is used for fraud detection.
When a user signs up to the platform, the relevant information is sent to the Amazon Fraud Detector API endpoint.
Amazon Fraud Detector generates a risk score (in the range of 0–1,000) on the input data — with 0 representing the lowest possible risk, and 1,000 indicating the highest possible risk. The detector returns the outcome based on the business rules: “allow”, “monitor”, or “block”.
The “allow”, “monitor”, or “block” results, including decisions by the support team, are stored in Amazon S3 for future model retraining and improvement.

After implementing this fraud detection solution, the ICONY support team saw the time they spent dealing with fake and spam accounts fall by 77%. This freed up the team to deal with individual user checks, which immediately improved quality on the platform and caused fraud reports from the community to drop by 63%. Moreover, the number of registered users returning to the platform has increased 4.13%. With less harassment from fake accounts and scammers, users feel more comfortable on the platform and enjoy using it.

These numbers highlight just how effective Amazon Fraud Detector can be in reducing the burden on the support team and improving the overall user experience, allowing ICONY to provide the best possible platform to its business partners. We hope this will inspire other teams to explore machine learning-based fraud detection using Amazon Fraud Detector.

Uwe Thomas has been CEO of ICONY GmbH, based in southern Germany, since 2016. ICONY GmbH is a small company with 15 employees who work every day to provide users in the ICONY network with the best service and a lot of fun when looking for a partner. However, Uwe himself is married and has 2 children.

Anna Grüebler is a Specialist Solutions Architect in Artificial Intelligence at AWS. She has more than 13 years experience in developing and deploying machine learning projects both in academia and in businesses of all sizes in Venezuela, Japan, and the UK. Her passion is taking new technologies and putting them in the hands of everyone, and solving difficult problems leveraging the advantages of using AI in the cloud.

Quick Cost Optimization Strategies for Early Stage Startups

AWS Editorial Team — Thu, 09 Dec 2021 21:17:47 +0000

Cost optimization is a top of mind consideration for any startup and can be achieved with a wide variety of techniques, but how you tackle it depends on the stage of your business’s growth. Unlike enterprise companies, startups are laser-focused on product development. This can force startups to choose between time spent building extra functionality to manage costs, like reorganizing account structures or building cost analytics pipelines, and prioritizing low-effort-to-high-impact architectural changes to keep your momentum up. In this post, we’ll share three easy-to-implement cost optimization strategies to help you quickly understand and optimize your spend, then get back to building features that will drive value for your customers. The three main concepts to focus on are spend awareness, architecture adjustments, and usage discounts.

Spend Awareness

Before making any changes, it’s important to understand what and where you’re currently spending. AWS Cost Explorer and AWS Budgets are the most efficient tools to help you make informed cost decisions.

Understand the Trends

AWS Cost Explorer helps you understand general spending trends quickly. As your business scales, so will your infrastructure costs, and Cost Explorer gives you a bird’s-eye view of that spend. The home view of Cost Explorer shows your last six months of spend grouped by service, which will help you assess the following:

Where should I start my cost optimization efforts? The default view will show the top five services contributing to monthly spend, which is helpful if you’re wondering which service you should start cost-optimizing and looking to make the biggest impact on spend.
Am I spending a significant amount on a service that doesn’t fit normal patterns? Your answer may be workload-dependent, but, for example, storage costs are typically lower than compute costs, so you may not expect EBS volume spend to be significantly higher than your spend on Amazon Elastic Compute Cloud (Amazon EC2).

Set a Budget

Even if you don’t have a fixed budget for your infrastructure spend, we still recommend implementing AWS Budgets to track costs and get alerts when you’ve reached certain spend thresholds. Accidental misconfigurations could result in a painful bill, and alerts can help you catch and remedy mistakes before they snowball. If you don’t have any budgeting alerts set up, take a few minutes to follow along with this tutorial to set one up.

Architecture Adjustment

Once you understand your costs, you’ll want to consider streamlining your architecture to be more cost-efficient. The changes we’ve highlighted below could yield the largest savings for the smallest amount of effort, so you can quickly get back to building your product.

Turn Off Resources When Not in Use

This advice may sound obvious, but the easiest, fastest, and most impactful way to achieve cost savings is to turn off resources when they’re not being used. In an average workweek, 70% of the hours are non-working hours. Imagine how much you could save by turning off non-production resources during that time! Tips to help you identify what to turn off:

Use tags: Use tags to determine what to turn off and differentiate between production and non-production resources.
Check Trusted Advisor: The Cost Optimization pillar of Trusted Advisor, as shown below, will show idle resources, and is a good starting point to check for resources that can be turned off if you haven’t tagged anything.
Identify non-constant workloads: Workloads that aren’t constantly running can be turned off and turned back on again when needed. For example, you can stop Amazon SageMaker notebooks when not in use. Amazon Redshift has a pause and resume feature to make this even easier.

Even if you know what needs to be turned off, that won’t yield savings unless you actually turn them off. If you struggle with this, automate the process by leveraging solutions like the AWS Instance Scheduler to configure start and stop schedules for Amazon EC2 and Amazon RDS.

Use the Newest Offerings

A simple way to reduce costs while maintaining performance is by using the latest and greatest of what AWS has to offer, including:

Using the latest generation of an instance type: Using the latest version of an instance type, for example, moving from m4 to m5, will improve price performance.
Exploring new instance types like the ARM-based AWS Graviton2 instances: Graviton2 processors offer up to 40% better price performance compared to current-generation x86-based instances. In addition to EC2, Graviton2 instance types are available for use with managed services like Amazon RDS, Amazon Aurora, Amazon ElastiCache, Amazon OpenSearch, and Amazon EMR. Because managed services eliminate infrastructure management tasks, switching to Graviton2 is a great way to yield cost savings without application code changes.

Usage Discounts

If you’re running a steady-state workload, you can confidently generate commitment numbers for a usage discount. But what if you’re in your early stages of growth, or have unpredictable usage? Compute Savings Plans help you strike a balance between evolving usage and getting savings for what you already use.

Why Compute Savings Plans and not EC2 Instance Savings Plans? What about Reserved Instances? Compute Savings Plans give you more flexibility than EC2 Instance Savings Plans and Reserved Instances, making them the ideal choice if you’re still in the process of making architectural changes. They offer a percent discount on any compute usage (Amazon EC2, AWS Fargate, and AWS Lambda), meaning you can always use the newest generation of hardware while still getting savings. It’s also an ideal option if you think you may switch instance types or re-architect between self-managed compute and serverless.
How big of a commitment should I make? The Savings Plan recommendations provided in AWS Cost Management give a baseline value of how much to commit to. Because Savings Plans are a dollar-per-hour commitment, the recommendation is sized against the minimum amount of compute you use every hour. You can stack multiple Savings Plans, so it’s good practice to start off with a conservative plan, and add more plans later if you continue to hit 100% use.

Conclusion

Cost optimization is a continuous process and should be a part of your software development lifecycle. These suggestions are only the beginning of how early stage startups should be thinking about the process of cost optimization. Your development teams should consider cost when deploying new features, and your AWS account team can also be a great source to guide you in through the optimization process process. By understanding your spend using AWS Cost Explorer and AWS budgets, turning off resources not in use, adopting the newest hardware, and leveraging Compute Savings Plans for flexible usage discounts, you can stay agile, while streamlining costs.

	Melissa Kwok is a Solutions Architect at AWS, where she helps customers of all sizes and verticals build cloud solutions according to best practices. When she’s not at her desk you can find her in the kitchen experimenting with new recipes or reading a cookbook.
	Faisal Farooq is Solutions Architect at AWS on the Startups team. He routinely hosts customer open forums to help Startups to discuss the industry wide challenges. In his prior role, he worked with Fortune 100 companies as a cybersecurity consultant. He is passionate about helping startups use AWS more efficiently and securely.

Built on AWS, Voltaiq Unlocks Battery-Powered Product Innovation at Amazon

AWS Editorial Team — Mon, 06 Dec 2021 21:24:09 +0000

Guest Post by Eli Leland, PhD, CTO and Co-Founder, Voltaiq

It’s a battery-powered world

Batteries are everywhere in today’s world. Much as microchips did through the 1980s and ‘90s, batteries are now finding their way into many devices and systems around us, and in the process are changing what we previously thought possible. Advances in lithium-ion battery technology in particular have enabled transformations in the way we connect and find information, the way we move people and goods, and the way we power our homes and buildings. These days, one might even go so far as to say that “every company is a battery company” in one way or another.

The transition to a battery-powered world is not without its challenges, however. We’ve all had the experience of our mobile phone running out of juice before lunchtime, and headline-grabbing stories of battery fires in peoples’ pockets and garages drive home the importance of making these products safe and reliable. Easier said than done when it comes to batteries.

While mature technologies like the mechanical components (springs, hinges, enclosures) and semiconductors that make up most of our devices are well understood and tend to fail in predictable ways, batteries are much more complicated. With myriad chemical and physical processes constantly unfolding inside them, and countless tiny ions “breathing” back and forth between the anode and cathode with each charge-discharge cycle, batteries in many ways are like living organisms. Because of this complexity, the only way to truly understand batteries, how they perform, how long they last, and how they fail, is to observe them under every conceivable operating condition, and take a detailed look at what you find.

Amazon as a battery leader

As a company known for innovative technology, Amazon has a few high-profile battery-powered product lines that have gained broad popularity with consumers, having introduced the Kindle e-reader in 2007, and later expanding to include products like Echo Buds, the Halo Band, and Amazon Astro. These products and many others are developed by a team within Amazon called Lab126. Initially created to develop the first Kindle, Lab126 has now expanded to develop a broad set of Amazon’s high-profile consumer electronics. And, naturally, a lot of those products have batteries in them. A lot.

It’s no surprise, then, that Lab126 has built a team tasked specifically with developing the battery power systems for these devices. It’s a big job that has a number of important facets. Because of the inherent complexity of batteries, battery system design includes steps to:

Choose an appropriate battery chemistry and cell design,
Identify a cell supplier that can deliver consistent quality,
Fully characterize the battery’s behavior across the full range of intended application scenarios,
And make sure that as long as the product is used as intended, the battery will meet the product’s warranty lifetime.

This process requires the team to test a large number of batteries (up to hundreds per new product line), and observe the detailed nuances of how each battery behaves as it is charged and discharged hundreds or even thousands of times. This work is done using specialized battery test equipment that produces large data files recording how the voltage and current rise and fall over each and every cycle, much like an EKG recording the battery’s “heartbeat.” With this quantity of battery testing and data analysis required for each product line, and product refresh cycles now being measured in months and not years, it quickly becomes a challenge to simply manage all of this data, let alone derive significant insights at scale.

Streamlining and accelerating product development with Voltaiq

As the Lab126 battery program grew, the team sought solutions to help them streamline, automate, and accelerate this work. They found one in 2016 when Lab126 began using Voltaiq, a cloud-based Enterprise Battery Intelligence (EBI) analytics software platform built on AWS, and offered by a Berkeley, California-based company of the same name. Founded in 2012 by two veteran energy and software entrepreneurs, Voltaiq works with a global customer base spanning transportation, consumer electronics, energy storage, and the full battery supply chain, helping them to launch products faster, optimize performance and reliability, and minimize risks from warranty returns and recalls.

The Voltaiq platform automatically aggregates and analyzes large quantities of battery data, both time-series logs of voltage and current, as well as metadata describing how batteries are built and operated, to reveal the key insights and correlations that drive battery performance and related desired business outcomes. Returning to the EKG analogy, in the same way that a cardiologist can look at the waveform of a heartbeat and identify a murmur, Voltaiq analyzes current and voltage waveforms from a battery’s operation to automatically identify signs of degradation and failure — things like unwanted chemical reactions inside the battery, signs of contamination in the battery materials, or physical deterioration of the battery’s internal structure.

At Lab126, Voltaiq automates the collection and processing of all data from across the Lab126 battery testing operation, “harmonizes” the data to a common format regardless of source, and makes this data and analysis instantly available to anyone on the team, wherever they happen to be. Today, Voltaiq is used extensively at Lab126 for developing battery-powered products.

By enabling enterprise-wide visibility, optimization, and decision making, the Lab126 battery team can move much faster, with better insight and greater automation across their product lines. “Having a centralized hub for data is critical as battery projects become more complicated. Voltaiq helps us better visualize our battery data and enables us to build safer and more dependable devices,” said Denys Zhuo, Senior Battery Systems Engineer at Lab126.

Enabling resilience for battery-powered organizations

As the Covid pandemic struck in early 2020, Voltaiq demonstrated another layer of value to the Lab126 battery team. As developers of physical products, the team was accustomed to working day-to-day at the Lab126 facility in Sunnyvale, California — suddenly the team found themselves working primarily from home. It was fortunate, then, that the automation and cloud-based nature of the Voltaiq platform enabled the team to work remotely with almost no impact to testing. ”As our team has shifted to a more remote work environment, Voltaiq enabled us to continue working well together as a team, limit the impact to battery testing and data analysis, allowing us to deliver on our objectives without any time delay,” said Bryan Holmdahl, Senior Battery Design Integrity Engineer at Lab126.

As the global wave of electrification continues to gain momentum, an increasingly broad set of companies will find that they need to develop a core competency around batteries, much the same as they have had to around the internet, and around microchips before that. Many of these businesses will need to develop capability around testing and analyzing the batteries that power their products. Typically, however, battery teams are missing both the IT infrastructure and the staff needed to properly manage and analyze the data volumes they face. Additionally, these same battery teams are under enormous time pressure to make design decisions that will have significant technical and financial impacts on the products they are bringing to market. Fortunately for Lab126 and thousands of other “battery-powered businesses”, the AWS-powered Voltaiq Enterprise Battery Intelligence platform can help them accelerate product development as they scale, with the confidence that comes from data-driven insights and decisions.

Eli Leland is a Voltaiq co-founder and the company’s Chief Technical Officer. Prior to Voltaiq, Eli developed a new capacitor technology as lead engineer on a U.S Department of Energy (DOE) Advanced Research Projects Agency (ARPA-E) project at the CUNY Energy Institute. Before CUNY, Eli worked for Trilogy Software, where he successfully deployed large-scale enterprise software systems to the Fortune 500. He was a Mirzayan Policy Fellow in Energy and Environmental Systems at the National Academies in Washington, DC. Eli has an MS and PhD in mechanical engineering from UC Berkeley, and a BSE in mechanical and aerospace engineering from Princeton.

Koo App Connects Millions of Voices in Their Preferred Language with AWS

AWS Editorial Team — Wed, 01 Dec 2021 19:59:25 +0000

When the social media revolution began, e-commerce sites mostly catered to English speakers, which left out a huge population of would-be participants. Koo, a microblogging platform based in India, noted the lack of inclusivity and made it their mission to create an app that is accessible to the entire spectrum of languages spoken in India. “How do we give them a platform so that everyone is able to share their thoughts, their opinions, without any language barrier?” asks Phaneesh Gururaj, President of Technology at Koo.

Koo originally started with just three languages—English, Hindi, and Kannada—and expanded from there. “Just yesterday, we added four more languages,” says Gururaj. “We plan to cover most of the primary languages spoken in India, which will be over 20 as we scale up. And this is just India.” They have also added—and plan to continue adding—different Nigerian dialects.

For creators, the process is straightforward. After they make a piece of content, they have the option to translate it into the language of their choice. Koo automatically does the translation and the transliteration, then shows the creator a preview before they post it. “Once the content is posted, the content is able to reach those who speak many other languages,” says Gururaj.

One major hurdle for Koo is that their software must have a rich, complex understanding of each language in order to translate not just literal meaning, but also the sentiment beneath words and phrases. “We want to be a part of the journey where we enrich our own language models with the content that we generate and go into that virtuous cycle of ensuring the model accuracy also increases,” says Gururaj. While there’s no shortage of data when it comes to NLP (Natural Language Processing) for English, Koo must pioneer their own models for many of the more obscure, or less widely utilized, languages. Koo employs AWS’s machine learning orchestration platform, Amazon SageMaker, which they combine with their own machine learning services.

As they scaled to millions of users, they observed both performance and replica lag issues and turned to AWS’s relational database service, Amazon Aurora. Post their migration to Aurora, the customer observed 20% faster query performance, no replica lag, and better storage scaling experience. “I think AWS has been a great partner for us,” says Gururaj, “A lot of the team we have assembled in the last five, six months, they all have come with a good knowledge of AWS…So there’s a lot of collective intelligence.”

Koo also employed Amazon Elastic Kubernetes Service (EKS), which assists in creating, scaling, managing, and terminating nodes for their cluster, to help a customer team that was manually scaling their Amazon EC2 cluster to meet a massive surge in traffic. With AWS’s assistance, they were able to migrate their production, pre-production, and development environments to Amazon EKS leveraging EC2 Spot Instances. As a result, their EKS cluster was able to handle 100% of their current traffic and scale appropriately per user growth. “We have been very happy with a lot of the discussions that we have been having with the solution architects and the other teams at AWS to scale,” says Gururaj.

After winning the Aatmanirbhar Bharat App Innovation Challenge last year, a government-run search to locate the Indian apps most likely to scale, Koo was able to onboard both a creator and a consumer network. They have also been concentrating on adding more creator-friendly tools, including an industry-first, patent pending feature called Multilingual Koo.

Looking forward, Koo plans to keep scaling, which means adding more languages and catering to more users in various countries. In other words, becoming even more inclusive. Their current goal is to have 100 million users on the platform. “To get to that stage, we need to build our community, we need to scale our systems, we need to introduce more innovative features, we need to strengthen our language modeling, we need to bring about a much better experience for our users,” says Gururaj. “There’s a lot more ground to cover in India and Nigeria specifically at this point of time. And we want to take this platform to much larger audiences outside India and Nigeria as well.”

Too Good To Go: Saving the Planet One Saved Meal at a Time

AWS Editorial Team — Tue, 30 Nov 2021 17:37:07 +0000

Food waste is a global issue that stretches far beyond its impact on underfed populations: it’s one of the primary drivers behind the climate crisis, representing 10% of all greenhouse gas emissions. Combating food waste reduces methane emissions in particular, while also preventing the waste of the labor and resources required to make the food.

But there’s room for hope, and Too Good To Go is helping make it happen with their B2C app. Too Good To Go is a free app dedicated to reducing food waste worldwide by connecting customers in Europe and North America to restaurants and stores with surplus food, helping consumers save, businesses thrive, and contributing to sustainability efforts.

“It’s good for the environment, it’s good for consumers, and it’s good for the partners,” says Robert Christiansen, Vice President of Platform Engineering at Too Good To Go.

Founded in Denmark in 2015, Too Good To Go quickly expanded throughout some of the largest cities in Europe before coming to North America in October 2020. According to their official story, food waste had become so normalized that it became an almost accepted part of life.

“We simply could not continue to see 1/3 of food produced thrown away, and decided that something had to be done. Food is simply too good to go to waste, and our journey started with this strong certitude,” Christiansen explains.

Through the app, a potential customer reviews restaurants, bakeries, and other outlets that have surplus food within a certain radius and during a specified time period. If they see something they like, the customer can then pay through the app and pick up their order from the outlet at a deeply discounted rate, making meals extra affordable and saving food that otherwise might end up thrown away.

Helping Too Good To Go achieve its goals

Too Good To Go is working to combat food waste while making food more affordable. But they can’t do it alone. Whenever the mobile app has grown to serve customers across Europe and North America, AWS has been there to help Too Good to Go build and scale.

Since its launch, Too Good To Go has helped close to 50 million users save over 100 million meals through more than 130,000 cafés, restaurants, bakeries, hotels, and more. And while Too Good To Go has shown significant growth since its inception, the objective is still the same. “It’s about saving the last meal,” says Christiansen. “So, the difference is scale.”

Understandably, Too Good To Go’s reach was more modest than its ambitions during its first few years. “I think it was four countries, maybe five countries back then,” says Christiansen. “It was a much smaller team. And there were a lot of independent stores like bakeries, small restaurants, pizzerias. These days, we’re a lot bigger. We work with giant chains. We also work with very big producers of food—Nestlé, for instance. So, it’s just getting bigger, and as with any company that grows, things change. But the primary goal has always stayed the same: we’re still trying to save meals from the bin, basically.”

Safety as a global issue

Achieving their goals has proven complex, especially as different countries—and in the U.S., different states—have different rules about everything from commerce to what qualifies as food waste. The move across the Atlantic was particularly challenging, given that the United States reports higher instances of food insecurity nationally than Europe does. According to the Food and Agriculture Organization of the United Nations, the average rate of food insecurity in North American households was 17.28%, compared to 10.42% in European households.

And given the high rate of food insecurity in the U.S., the impact of food waste is significant—especially when the average American household wastes one-third of its purchased food according to the American Journal of Agricultural Economics. “Americans think about food waste a lot differently than Europeans,” says Christiansen. “We didn’t know food insecurity was such a big problem. And we want to help other food programs, such as food banks so that’s also something we are working on in the U.S.”

Coming to the United States also meant making significant adjustments to how the Too Good To Go app operates. “Of course, there are the technical things,” says Christiansen. “Going to the U.S., opening up multiple data centers…AWS will still work when you have a system that’s not built, and when I started, we had an entirely different system than we have now. We basically have changed everything before that to scale. And from the scaling point of view, we’re growing, so it’s not just one database.”

It was also important for Too Good To Go to keep user data safe across different continents. “The biggest infrastructure change we made was having multiple data centers for having full GDPR compliance and also having people’s data in their own region,” Christiansen says. “And that alone requires a lot more infrastructure to handle miscommunication, because as a consumer in the U.S., you can still go to Europe and buy from a European store—you’re not locked in the U.S. So all that requires a lot more moving parts.”

AWS has helped Too Good To Go grow organically

Too Good To Go CEO Mette Lykke

Because Too Good To Go is more of a connecting platform than a storefront, effectively juggling those moving parts is especially important. “Too Good To Go in itself does not sell food,” Christiansen says. “All the food is sold by the store or restaurant, and we simply facilitate it. Because when selling food, both in the U.S. and in Europe, and I guess all over the world, you need local entities, and there are a lot of local rules involved when it comes to food. Unlike Amazon, we don’t sell stuff ourselves on our own marketplace. And that’s a very important distinction. We make the connection between two parties, and we then take care of everything else for our partners and consumers.”

All of which has been built through AWS. Christiansen explained that when Too Good To Go started, they were using classic Amazon Elastic Compute Cloud (Amazon EC2), but as they’ve grown as a company, so have their needs. Often, it’s about flexibility on the fly.

“What changed is that we’re using more and more services because there are a lot more business needs,” Christiansen says. “Data is one of the areas where I can only see it going upward in complexity – then there’s the scale for processing data and figuring out how to work with that data. We’re using the stack at Amazon, and we’re using more services. And we especially like those services that can scale. We have traffic that jumps quite wildly when marketing decides to do big campaigns, and we have 10 times the traffic. So, to the extent we can, we use all these auto scaling services that are finely tuned.”

Oliver Meyer, a data engineer with Too Good To Go, says that the success of their objective is predicated on being able to adapt quickly and seamlessly. “The amount of demand of data coming in is growing pretty fast,” says Meyer. “And the variety of data, the variety of use cases of people accessing it, of ways in which they’re accessing it, everything is growing. And then with that, the infrastructure has to follow.”

Rising to those challenges isn’t just an essential component of what makes Too Good To Go work—it can also be energizing. “I think that applying new services or new features to new problems is something that’s very, very motivating,” says Christiansen. “And sort of this idea of facing a new challenge, and then digging into what we have and what we can add to handle this, it’s very broad, and it’s a bit abstract, but that’s really a process that I enjoy. And AWS is a great provider to work with in terms of that, simply because of the variety of services. So, any time a new problem or a new challenge comes up, it’s always quite fun to dig into what we can throw at it and see what might work.”

8 Reasons to Attend AWS re:Invent 2021 Virtually

Keely O'Neill — Mon, 22 Nov 2021 21:32:07 +0000

As AWS re:Invent 2021 nears, many attendees are packing their blazers and sensible sneakers to head to Las Vegas. But if you’re not able to be there in person—or if you’re allergic to neon signs and magic shows—here are eight reasons why you’ll get just as much out of attending virtual re:Invent.

1. Virtual re:Invent is free, which means you can take the information and inspiration without the gamble of buying a ticket.

2. Sometimes attending an event online means missing out on live perks. But at virtual re:Invent, online attendees will be able to participate in the same live leadership and keynote sessions, plus all of the breakout sessions.

3. AWS works hard to create accessible events, and we are implementing features like closed captions in English, as well as translations for keynotes and leadership sessions in eight different languages.

4. We’re big proponents of the follow-the-sun model, which encourages employees to work when it’s most convenient for them. Similarly, all of the virtual re:Invent sessions will be streamed using the same model. This means you can watch “live” from the comfort of your home from anywhere in the world—regardless of geographic location or time zone.

5. If any questions pop up while you’re watching, you can access Ask an Expert virtually. These live one-on-one chats will enable virtual attendees to connect with AWS experts to get answers to technology and business questions in real-time.

6. Still have FOMO? Build community with fellow online attendees, through three live, shared experiences that will stream daily:

Live from re:Invent is a broadcast series that will highlight three main areas: a behind-the-scenes peek into what it takes to make re:Invent come together, on-the-street community reporting, and featured partner segments. Catch these shows by tuning into the dedicated “Live from re:Invent” channel. Each will run at the top of every hour from 8:00 AM – 4:15 PM PST. If you’re looking to watch these shows later, they will be available on demand via the event platform on Monday, December 6.
For our international community, AWS on Air will be recapping re:Invent highlights in non-English languages, including Chinese, Korean, Italian, German, French, Portuguese, and Spanish. These nightly time slots will simulcast to the virtual event platform via a dedicated channel.
theCUBE, which is known for covering enterprise technology thought leadership around the globe, will offer virtual attendees live coverage from re:Invent. Online participants will be able to access theCUBE’s coverage via a dedicated channel page within the virtual event platform.

7. Rev your engines, because virtual attendees can also watch the world’s first global autonomous racing league, AWS DeepRacer, which will take place before re:Invent starts. These races are open to developers of all skill levels, who can submit their models and compete for both prizes and good, old-fashioned glory! Hosted by AWS DeepRacer, these competitions will also be recorded and available to watch on-demand.

8. Lastly, virtual attendees can participate in Partner Solutions Day (PSD), a networking event that will take place entirely within the virtual re:Invent platform. PSD will be held on Monday, December 6, the first day of week two of the virtual experience. Take your next step in your cloud journey by pairing up with an implementation partner, learning how to best leverage AWS technologies, and discovering how the AWS Partner Network can help you build, market, and sell your AWS-based offerings.

Returning to re:Invent in Las Vegas this year will be exciting, but you’ve also got eight good reasons to register now and enjoy the jam-packed schedule and custom experiences of virtual re:Invent from the comfort of your own home. After all, as they say in Vegas, the house always wins.

How claimsforce Built a Future-Proof Lake House with AWS

AWS Editorial Team — Fri, 19 Nov 2021 15:41:29 +0000

Guest post by Robert Kossendey, Team Lead Data Chapter, Johannes Kotsch, Data Engineer, and Hans Hartmann, Data Engineer, of claimsforce

In July 2021, Germany and several other European countries were hit by a 100-year flood, which devastated property and killed more than 180 people. As a result, our partner insurers and damage adjusters received an unforeseen number of claims in a short amount of time from people in dire need of help.

At claimsforce, this spike in demand resulted in a 300% load increase on our systems, but we had no issues scaling our infrastructure. AWS’s seamless scalability meant that we could continue operating with no downtime or maintenance effort, which enabled us to focus fully on the response to the natural disaster, providing our partners with valuable insights, and offering assistance to the people whose lives had been affected by this horrible tragedy.

This catastrophic flooding is a striking example of how claimsforce can have meaningful real-world impact. As an InsurTech startup based in Hamburg, Germany, our mission is to create great claims experiences. Our products enable insurers and damage adjusters to process claims more efficiently by digitally enhancing their processes, as well as the customer experience. This makes us one of the key players in the digital transformation of the European insurance industry. We differentiate ourselves by using data to tackle difficult problems like fleet routing or recommendations and maximize the experience for our customers.

As a fast-growing startup, we deliver products to multiple insurers and damage adjuster networks, which leads to an ever-increasing amount of data stored in our application databases and external sources. Initially, we had no dedicated infrastructure in place to process the data. But performing tasks on distributed data has become harder, and the growing volume of data provides great opportunities to gain insights that will benefit our customers.

Why AWS?

We decided to build our data infrastructure on AWS because it offers serverless and fully-managed services. In contrast to traditional products, AWS services require no additional administrative effort on our part, giving us the ability to focus fully on data analysis and product development—something that has been especially valuable to us as a startup with limited resources. We also appreciate the seamless scalability and pay-per-use cost structure. At the moment, we use a relatively small amount of data compared to large enterprises, so we benefit from the flat cost structure of AWS services. However, we are also able to scale for the future and account for load peaks without a problem.

Choosing a Lake House architecture

We chose the Lake House architecture for our data infrastructure because we think it combines the best of both worlds from data lakes and data warehouses.

We regularly face rapid changes of requirements, leading to changes in available data and its structure. A data warehouse with a static schema was therefore not an option for us. Meanwhile, a pure data lake lacks analytic capabilities and can easily become a “data swamp” through a lack of schema enforcement, metadata management, and documentation.

AWS offers a range of services for implementing a Lake House, which can be integrated smoothly. We use Amazon Simple Storage (Amazon S3) as a standardized object store for data. AWS Glue is the go-to service for all kinds of data migration and ETLs. At the same time, AWS Glue provides a metadata store that documents and helps us enforce a schema on our object data. That process is very convenient and requires no update maintenance. The data can be queried directly using the Amazon Athena query engine, providing an ideal “single source of truth.”

Our architecture

Our main data sources are our Amazon DynamoDB application databases, external APIs, and user tracking data. We use AWS Lambda functions as a universal and adaptive way to gather data and write it to a specific Amazon S3 bucket. The S3 bucket provides a landing area for the raw data, with no transformations to eliminate any possible data loss. This also allows for future changes in analytics requirements, as the original data remains intact.

This raw data gets crawled by the AWS Glue crawler, scheduled daily. The crawler determines the schema of all our data and persists the schema in the AWS Glue data catalog. Changes to the schema get detected and updated at each crawler execution.

Cleansing in AWS Glue

The raw data from S3 is processed with several cleansing jobs in AWS Glue. These cleansing jobs use the data definitions in the AWS Glue data catalog as descriptions for the input data. During the cleaning process, we enforce different data standards, which we have defined together with all stakeholders, such as business analysts. These data standards include:

Unified date format and time zone
Standardized column-naming conventions (e.g., snake case)
Handling of empty values (null instead of empty strings)
Transformation of complex data types (arrays, maps, JSON-strings) into atomic data types

After enforcing these data standards, our cleansing jobs write the data into another S3 bucket that contains our processed data. The schema of our processed tables gets automatically updated in the AWS Glue data catalog by the AWS Glue job. To prevent our jobs from processing the same data twice, we are using the job bookmark feature that AWS Glue provides, which enables AWS Glue to compute only the data that was added since the last job execution.

We use Athena to perform SQL queries on data saved in our two S3 buckets. Athena is very well-suited to ad hoc requests because it can perform SQL queries directly on S3—no additional data moving to a data warehouse is needed. By keeping our metadata up to date using the AWS Glue data catalog, we always have a defined schema in our data buckets. Our data analysts and data scientists can perform data exploration for ad hoc requests, even on dark data.

We store data in the Parquet file format, a columnar storage format. This allows us to process and query data more efficiently, because not every row of data needs to be queried to get a result.

Since every cleansing job results in a single file, and because we run our cleansing jobs on an hourly basis, we end up with a relatively high number of files per day and an increased number of file reads from S3. To reduce the number of S3 get-requests (and costs), we wrote a Parquet compaction job, running on a daily basis. The job compacts all of our daily files, so they adhere to the most efficient Parquet file size, 512 MB to 1024 MB.

Our two use cases for data: business analytics and machine learning

For business analytics, a use case we see frequently is around the need to support our customers with valuable insights about the efficient use of their resources. For that, we use Amazon’s fully managed data warehouse, Amazon Redshift. This allows us to perform efficient SQL queries on large amounts of data.

The Redshift data is further aggregated to account for business-level requirements. We achieve these aggregations with our AWS Glue aggregation jobs. They read from multiple tables in the processed bucket, with the schemas defined in the AWS Glue data catalog, and perform the needed operations like aggregating or joining.

With those analytical capabilities, we cluster past claims by region and search for regions with many claims per damage adjuster, long response times or high rejection rates from damage adjusters. With the help of those insights, our partners can identify regions where more resources are needed to improve their customer experience.

We also use the data in Amazon Redshift as the basis for our visualization tools, Tableau and Amazon QuickSight. Those tools provide dashboards and analyses for our customers and for internal use.

In another use case, we use Amazon SageMaker, Amazon’s fully managed machine learning platform, to build our disposition engine. This feature recommends the best damage adjuster for a specific claim. We used historical data from claim assignments that we could directly extract from the processed S3-Bucket. Having Parquet as a standardized file format made it easier to ingest large data into our ML-Models than querying large amounts of data via SQL.

The use of our disposition engine resulted in better assignment decisions, using resources more efficiently. With the help of our predictions, driving times of damage adjusters could be reduced, and their workload could be distributed more equally.

Conclusion

Ultimately, our decision to build a Lake House with AWS has offered us clear benefits. For one, storing all ingested data in a raw bucket paid off. There were some cases where we needed to change the data structure in the processed bucket, but with all the raw data in place, we could just change the transformation logic, delete the data in the bucket, reset the job bookmark, and execute the AWS Glue job again. Moreover, the AWS Glue jobs have no difficulty processing large amounts of data. This is a huge advantage compared to other services that could be considered for data moving.

Overall, even after just a couple of months of using our Lake House, it’s clear that it’s had a positive impact on our business, allowing us to improve our efficiency and flexibility and focus on delivering the best possible customer experience. That’s especially important when dealing with stressful circumstances like a flood or other natural disaster, when speed, convenience, and reliability are key.

	Robert Kossendey is the Team Lead Data Chapter at claimsforce. His focus areas are highly distributed cloud architectures and big data processing. If there is some free time, Robert likes to dive into Data Science and Machine Learning topics.
	Johannes Kotsch is working as a Data Engineer in our disposition product team, which is in charge of matching the best damage adjustor to a claim. He is passionate about DynamoDB and everything related to distributed data in the cloud. Currently, he is also pursuing his Master’s degree in Computer Science with a focus on Big Data.
	Hans Hartmann is a Data Engineer in one of our product teams, responsible for the claim assessment. In addition he is currently doing his masters in Data Science. He’s deeply passionate about everything related to data.

Enjoy AWS re:Invent 2021 and More at the AWS Startup Loft in San Francisco

AWS Editorial Team — Thu, 18 Nov 2021 15:40:08 +0000

What happens in Vegas isn’t staying in Vegas—at least not when it comes to AWS re:Invent 2021. AWS is bringing the best startup content of the event (plus the party!) to the newly reopened AWS Startup Loft on San Francisco’s Market Street.

From November 29 to Dec 1, the Loft will be home to livestreams of keynote speeches from AWS leaders. But this is more than just a watch party. Check out everything you’ll be able to enjoy at the free San Francisco event:

Up close and personal with other founders: Get your questions answered in live breakout sessions featuring AWS Startup Solutions Architects and partners covering both technical insights and business fundamentals for startups that you can only find at the Loft.
Your most complex questions, answered: Come prepared with your toughest tech issues, plus any questions you have about funding, leading, or scaling your business. AWS experts will be on hand to answer them at your request, providing the personalized solutions that will carry your business to the next level.
Keynote livestreams: View high-quality livestreams from AWS leaders, including Adam Selipsky, CEO of AWS; Peter DeSantis, SVP Utility Computing and Apps; and Swami Sivasubramanian, VP Amazon Machine Learning.
Insight from more startup leaders than ever before: The Loft is committed to creating more inclusive spaces and better opportunities for traditionally underrepresented founders. This year’s diverse group of speakers will be a dynamic reflection of the wide variety of backgrounds and experiences that make up the San Francisco startup community, providing attendees with more comprehensive and forward-thinking insight than ever before.
The chance to unwind: Vegas won’t be hosting the only party. Relax with DJ breaks, trivia games, lunches, beer tastings, LED ping pong tables, happy hours, a dark room photo booth with a light artist, and other fun surprises (and swag!) throughout the three-day event.
Unparalleled networking: We’re looking forward to providing the space for founders to connect in person once again during the Loft’s inaugural event. Whether it’s through competing at happy-hour trivia or chatting over lunch about a great keynote, you’ll have the chance to form the partnerships that could define your business going forward.
Coworking resources: In case you need to get a little work done in between keynotes and ping pong, the Loft’s industry-leading coworking space will be open for business.

This is also your chance to get an exclusive sneak peek at some of the new resources in the re-imagined Loft. During the pandemic closure, AWS radically transformed the Loft from a coworking and event space to an always-on incubator—the hub you need to connect with fellow founders and accelerate your business.

The space will now house on-site resources like a state-of-the-art podcast studio and technical and business experts, plus offer hands-on programming that represents the reality of our diverse San Francisco startup community.

During re:Invent @ the Loft, all founders are invited to check out what the space has to offer, including you. We’re looking forward to welcoming entrepreneurs from any industry background and at any stage of your business, whether you’re a first-time founder in need of a network or a late-stage startup looking for custom advice on boosting your proficiency in core AWS services.

As the Loft reopens its doors, we are eager to protect the health and safety of all our guests. This event will require proof of vaccination, and attendees will be asked to wear a mask if they are not eating or drinking. Click here for more information on our health protocols.

Head over to our event and FAQ page to see the full schedule, meet speakers, and register today!

Monitoring SageMaker ML Models with WhyLabs

AWS Editorial Team — Wed, 17 Nov 2021 17:58:17 +0000

Guest post written by Danny D. Leybzon, WhyLabs MLOps Architect and Ganapathi Krishnamoorthi, AWS Solutions Architect

As the real-world changes, machine learning models degrade in their ability to accurately represent it, resulting in model performance degradation. That’s why it’s important for data scientists and machine learning engineers to support models with tools that provide ML monitoring and observability, thereby preventing that performance degradation. In this blog post, we will dive into the WhyLabs AI Observatory, a data and ML monitoring and observability platform, and show how it complements Amazon SageMaker.

Amazon SageMaker is incredibly powerful for training and deploying machine learning models at scale. WhyLabs allows you to monitor and observe your machine learning model, ensuring that it doesn’t suffer from performance degradation and continues to provide value to your business. Today, we’re going to demonstrate how to use WhyLabs to identify training-serving skew in a computer vision example for a model trained and deployed with SageMaker. WhyLabs is unique in its ability to monitor computer vision models and image data; whylogs library is able to extract features and metadata from images, as described in Detecting Semantic Drift within Image Data. The ability to create profiles based on images means that users can identify differences between training data and serving data and understand whether they need to retrain their models.

To follow along with this tutorial, we’ll utilize the WhyLabs AI Observatory, which you can access through the WhyLabs free, self-serve Starter tier.

Prepare the data and profile the training dataset with whylogs

The WhyLabs AI Observatory helps data scientists and machine learning engineers prevent model performance degradation by allowing them to monitor machine learning models, the data being fed to them, and the inferences they generate. It leverages the open-source library whylogs to generate “data profiles” (statistical summaries of datasets); these profiles are then sent to the WhyLabs platform, which allows users to analyze and alert on changes in these profiles over time.

For this example, we will build a simple model using SageMaker’s built-in algorithm for multi-label Image Classification. We will leverage whylogs’ unique ability to profile image data for data profiling and send the profiles to the WhyLabs platform for monitoring and observability. We are leveraging pycocotools 2017 images dataset for this example. This is a huge dataset with over 80 classes of images. For ease of use, we will consider only four classes and profile one image per class in the training dataset and send it to the WhyLabs platform.

The model in this blog was built using Python3 in an AWS environment on a Jupyter notebook instance. You can also run it in a Jupyter notebook on a local machine configured to run Amazon SageMaker training jobs. You will need an instance with Python3 and other Python libraries like SageMaker and WhyLabs used in the code. The dataset that we use is closer to 1GB, so make sure you have enough space in your notebook instance to download and process this dataset.

The IAM role attached to the notebook should have read and write permissions to Amazon Simple Storage Service (Amazon S3), and it should also have the access to launch training jobs and deployment endpoints in SageMaker. If you are running your notebook outside of AWS, you may want to ensure that you assume an IAM role that has these permissions attached.

Let this dissect the code to understand the flow.

First, we need to download pycocotools and install Python libraries using pip install.

import sys
!{sys.executable} -m pip install pycocotools
!pip install seaborn --upgrade
!pip install whylogs==0.6.8.dev0

Next, we will need to import certain libraries, including the SageMaker library for model training and deployment, the whylogs library for data profiling, and some standard Python libraries for numeric and image data manipulation.

import sagemaker
from sagemaker import get_execution_role
import whylogs
from whylogs import get_or_create_session
import numpy as np
from PIL import Image
import os
import json
import urllib.request
from pycocotools.coco import COCO
import datetime
import pandas as pd
from whylogs.app import Session
from whylogs.app.writers import WhyLabsWriter
 
role = get_execution_role()
print(role)
 
sess = sagemaker.Session()
bucket = sess.default_bucket()
prefix = "ic-multilabel"
 
print("using bucket %s" % bucket)

In order to utilize WhyLabs, we need to specify an API key and org ID as environment variables. To generate your own API key and org ID, follow this documentation for generating the key and ID using your free WhyLabs account.

# Set WhyLabs config variables
WHYLABS_API_KEY = "api.KEY"
WHYLABS_DEFAULT_ORG_ID = "org-12345"
 
# Set environment variables
os.environ["WHYLABS_API_KEY"] = WHYLABS_API_KEY
os.environ["WHYLABS_DEFAULT_ORG_ID"] = WHYLABS_DEFAULT_ORG_ID
 
today = datetime.datetime.now()
yesterday = today - datetime.timedelta(days=1)

We will use SageMaker’s built-in algorithm for image classification by referencing the latest container image as below

training_image = sagemaker.image_uris.retrieve(
 	region=sess.boto_region_name, framework="image-classification", version="latest"
)

Next, we’ll download the images from pycocotools. In our example, we will download a 2017 version of the image repository called ‘val2017,’ which also comes with an annotations json file.

def download(url):
 	filename = url.split("/")[-1]
 	if not os.path.exists(filename):
 	urllib.request.urlretrieve(url, filename)
 
# MSCOCO validation image files
download("http://images.cocodataset.org/zips/val2017.zip")
download("http://images.cocodataset.org/annotations/annotations_trainval2017.zip")

As a standard practice, we will unzip the file to access the underlying images and annotations file.

!unzip -qo val2017.zip
!unzip -qo annotations_trainval2017.zip

To make it simple, we will choose the first four classes in pycocotools images for extraction. The category ids are mapped as below:

Person
Bicycle
Car
Motorcycle

We will use pycocotools built-in libraries to extract images from these four classes by referring to the annotations and creating a lst file. A lst file is a text file that consists of a list of pointers to images in other files. The below code snippet gets the four image categories listed above and extracts the images within each category. A lst file is constructed with the image ids and the labels as shown below.

annFile = "./annotations/instances_val2017.json"
coco = COCO(annFile)
catIds = coco.getCatIds()
 
image_ids_of_cats = []
 
for cat in catIds:
 	image_ids_of_cats.append(coco.getImgIds(catIds=cat))
 
image_ids = []
labels = []
# use only the first 4 classes
# obtain image ids and labels for images with these 4 classes
 
cats = [1,2,3,4]
 
for ind_cat in cats:
 	for image_id in image_ids_of_cats[ind_cat - 1]:
 	if image_id in image_ids:
       	labels[image_ids.index(image_id)][ind_cat - 1] = 1
 	else:
       	image_ids.append(image_id)
       	labels.append(np.zeros(len(cats), dtype=np.int))
       	labels[-1][ind_cat - 1] = 1
       	 
# Construct the lst file from the image ids and labels
# The first column is the image index, the last is the image filename
# and the second to last but one are the labels
with open("image.lst", "w") as fp:
 	sum_labels = labels[0]
    
 	for ind, image_id in enumerate(image_ids):
 	coco_img = coco.loadImgs(image_id)
 	image_path = os.path.join(coco_img[0]["file_name"])
 	label_h = labels[ind]
 	sum_labels += label_h
 	fp.write(str(ind) + "\t")
 	for j in label_h:
       	fp.write(str(j) + "\t")
 	fp.write(image_path)
 	fp.write("\n")
       	 
 	fp.close()
print(sum_labels)

Now that we’ve extracted our training data, we can profile it and send those profiles to WhyLabs in a few easy lines of code.

writer = WhyLabsWriter("", formats=[],)
session = Session(project="demo-project", pipeline="pipeline-id", writers=[writer])
with session.logger(tags={"datasetId": "model-1"},dataset_timestamp = yesterday) as ylog:
 	# Profiling a sample image for person
 	ylog.log_image("val2017/000000066231.jpg")
 	# Profiling a sample image for bicycle
 	ylog.log_image("val2017/000000169996.jpg")
 	# Profiling a sample image for car
 	ylog.log_image("val2017/000000204871.jpg")
 	# Profiling a sample image for motorcycle
 	ylog.log_image("val2017/000000136715.jpg")

Profiling these images should take less than a second. From here, we can navigate back to the WhyLabs UI (where we had previously generated our API key) and head to the Model Dashboard to verify that our model is now populated with these profiles.

As we can see, our “Image Classifier” model is now populated with some profiles. We can explore this “baseline” training profile by navigating to the model summary page and then drilling down to that particular profile.

As illustrated above, whylogs was able to extract some key metadata (such as the pixel width and pixel height) and image features (such as the brightness mean and the standard deviation within the saturation) from the images we fed it. One particularly interesting piece of metadata is the ImagePixelWidth, which we will return to after uploading a profile from our inference dataset that doesn’t match the training dataset.

Train and deploy the model with Amazon SageMaker

Let us now move on to our standard practice of creating the lst file and splitting the input dataset into train and validation channels for the image file and the lst file.

!shuf image.lst > im.lst
!head -n 2500 im.lst > mscocoval2017train.lst
!tail -n +2501 im.lst > mscocoval2017val.lst
!head mscocoval2017train.lst
!wc -l mscocoval2017train.lst
!wc -l mscocoval2017val.lst
 
# Four channels: train, validation, train_lst, and validation_lst
s3train = "s3://{}/{}/train/".format(bucket, prefix)
s3validation = "s3://{}/{}/validation/".format(bucket, prefix)
s3train_lst = "s3://{}/{}/train_lst/".format(bucket, prefix)
s3validation_lst = "s3://{}/{}/validation_lst/".format(bucket, prefix)
 
# upload the image files to train and validation channels
!aws s3 cp val2017 $s3train --recursive --quiet
!aws s3 cp val2017 $s3validation --recursive --quiet
 
# upload the lst files to train_lst and validation_lst channels
!aws s3 cp mscocoval2017train.lst $s3train_lst --quiet
!aws s3 cp mscocoval2017val.lst $s3validation_lst --quiet

We now configure the SageMaker training job with required parameters including the container image for object detection algorithm, IAM role, type of training instances, number of training instances, maximum training time in seconds, EBS volume storage, input mode and output location of the model artifact. For deep learning, we normally recommend GPU instance families like P2, P3 etc. as training is much faster on a GPU instance, but if you are on a budget or attempting to explore and learn, you can use CPU instances like C5, C4 and C3. In the below code we are using a GPU instance for training.

s3_output_location = "s3://{}/{}/output".format(bucket, prefix)
multilabel_ic = sagemaker.estimator.Estimator(
 	training_image,
 	role,
 	instance_count=1,
 	instance_type="ml.p2.xlarge",
 	volume_size=50,
 	max_run=360000,
 	input_mode="File",
 	output_path=s3_output_location,
 	sagemaker_session=sess,
)

With any SageMaker training, it’s important to set hyperparameters for better accuracy. Every built-in algorithm would need a certain set of hyperparameters (required and optional) to be tuned as listed in the documentation. Here we set some static values to hyperparameters, but you can also pass on a range of values in a json document as shown here.

multilabel_ic.set_hyperparameters(
 	num_layers=18,
 	use_pretrained_model=1,
 	image_shape="3,224,224",
 	num_classes=5,
 	mini_batch_size=128,
 	resize=256,
 	epochs=5,
 	learning_rate=0.001,
 	num_training_samples=2500,
 	use_weighted_loss=1,
 	augmentation_type="crop_color_transform",
 	precision_dtype="float32",
 	multi_label=1,
)

We then set up four different channels for the SageMaker training job as below and kick off a training job.

train_data = sagemaker.inputs.TrainingInput(
 	s3train,
 	distribution="FullyReplicated",
 	content_type="application/x-image",
 	s3_data_type="S3Prefix",
)
validation_data = sagemaker.inputs.TrainingInput(
 	s3validation,
 	distribution="FullyReplicated",
 	content_type="application/x-image",
 	s3_data_type="S3Prefix",
)
train_data_lst = sagemaker.inputs.TrainingInput(
 	s3train_lst,
 	distribution="FullyReplicated",
 	content_type="application/x-image",
 	s3_data_type="S3Prefix",
)
validation_data_lst = sagemaker.inputs.TrainingInput(
 	s3validation_lst,
 	distribution="FullyReplicated",
 	content_type="application/x-image",
 	s3_data_type="S3Prefix",
)
data_channels = {
 	"train": train_data,
 	"validation": validation_data,
 	"train_lst": train_data_lst,
 	"validation_lst": validation_data_lst,
}

We now kick off the SageMaker training job that we just configured with a fit command.

multilabel_ic.fit(inputs=data_channels, logs=True)

Once the training completes, we deploy the model to a real-time SageMaker endpoint, so we can start inferencing. With this deployment, SageMaker spins up server(s) for inference based on the configuration, deploys the model artifact and processes inference payloads. We go on to configure instance type, instance count and serializer library for the SageMaker deployment endpoint as shown below. You can also refer to the recommended instance types for built-in algorithms here.

from sagemaker.serializers import IdentitySerializer
 
ic_classifier = multilabel_ic.deploy(
 	initial_instance_count=1,
 	instance_type="ml.m4.xlarge",
 	serializer=IdentitySerializer(content_type="application/x-image"),
)

The model is now ready for inference.

Make a prediction on out-of-distribution data

We will now showcase how WhyLabs can be effective in this scenario by taking an image from a class outside of what the model was trained on and lowering the image resolution. This realistically represents training-serving skew, where the data used to build the model is different from the data that the model tries to infer against in production. We can then use WhyLabs to analyze the inference data and see how it differs from the training data.

Let’s pick an image from a different class outside of what the model was trained and lower the image resolution.

## Profiling bird image
img = Image.open('val2017/000000147498.jpg')
print(img.size)
 
red_res_image = img.resize((200, 120))
red_res_image.save('low_res_image.jpg')

Now, let’s generate a whylogs profile of that image and send it to WhyLabs.

writer = WhyLabsWriter("", formats=[],)
session = Session(project="demo-project", pipeline="pipeline-id", writers=[writer])
with session.logger(tags={"datasetId": "model-1"},dataset_timestamp = today) as ylog:
 	# Profiling the inference image with reduced resolution
 	ylog.log_image("low_res_image.jpg")

In the WhyLabs UI, we can now see that we have two profiles for this model: one for training and one for inference:

We can see that our new profile has the same features as the old one:

And compare the two:

Looking at the ImagePixelWidth feature specifically, we can observe that the width of the images in first profile is significantly higher than in the second profile, indicating that our inference data is lower in resolution.

In this exercise, we’ve used a very simple example, and we knew that there would be some disparity (training-serving skew) because we set the training and inference data to be different. In the real world, this would be a clear indicator that our model might not perform as well in production as it did during our initial testing in our training sandbox. As an action to such an indication, we might need to take into account that lower image resolution leads to lower model accuracy and potentially retrain our model to help mitigate that.

Conclusion

Following the steps above, we trained a computer vision model in SageMaker and monitored the input image data with WhyLabs. We intentionally created synthetic data for production that didn’t match the data being used for training the model, so that we could highlight WhyLabs’ ability to identify training-serving skew. In the real world, data scientists don’t intentionally make their production data mismatch their training data. However, data in the real world tend to be messier and more prone to produce errors than those used by data scientists during the training and validation phase.

Training-service skew is one of a number of different examples of problems that can cause model performance degradation in production environments, but through this example, we saw how WhyLabs can help reduce this risk in image data. For those working with other formats, text support is available now, and video and speech data support will be coming soon.

To get support for the open source whylogs library, the self-serve AI Observatory, or for any ML monitoring questions, join the WhyLabs Community Slack.

	Danny D. Leybzon is an MLOps Architect at WhyLabs. Danny evangelizes machine learning best practices and has helped countless customers bring their ML models out of the sandbox and into production. When Danny isn’t thinking, talking, or writing about MLOps, he’s hiking, skiing, or scuba diving.
	Ganapathi Krishnamoorthi is a Senior Solutions Architect at AWS. Ganapathi provides prescriptive guidance to enterprise customers helping them to design and deploy cloud applications at scale. He is passionate about machine learning and is focused on helping customers leverage AI/ML for their business outcomes. When not at work, he enjoys exploring outdoors and listening to music.

Customer Focus Turned 1Password into a $2 Billion Business with AWS

AWS Editorial Team — Tue, 09 Nov 2021 16:54:01 +0000

Back in 2005, Roustem Karimov and Dave Teare were web consultants helping other folks build e-commerce sites when they started a side project to help keep track of all the different passwords needed for work. At the time, Roustem recalls thinking, “Hey, we’ll spend a few weeks on this, and then we’ll just go back to doing our real job web consulting stuff.”

But it became clear shortly after they finished building the product and put a purchase form up that Roustem and Dave wouldn’t be returning to their day jobs any time soon. “The very first hour we put it online, someone bought it,” Roustem says. Revenues from the first year totaled roughly $80,000, causing the pair to drop everything and devote their full attention to building 1Password. That turned out to be the right instinct: this past summer, the company raised a Series B round that doubled its valuation, making 1Password worth $2 billion today.

Founder Roustem Karimov

Roustem is quick to note, though, that it took a decade and a half to build the company to what it is today—and those early years were lean ones. “We lived on ramen noodles.” It was just Roustem and Dave, a two-person team, doing everything themselves: development, marketing, customer support, making sure the web store stayed online, “because if this thing goes down, you don’t have any revenue coming in. You’re basically out of business.”

He still remembers the panicked texts coming in at 2 a.m. “The service is down, wake up!” But that all began to change when 1Password started relying on AWS to manage their service. “It’s off my shoulders. Everything is taken care of for me,” Roustem remembers marveling. “The backups are happening, upgrades are happening—I don’t have to worry about that.”

Today, 1Password still has objects stored on Amazon S3 that date all the way back to 2007. “It may be the biggest compliment that you don’t really have to think about it too much,” Roustem says of S3. “You put things there, they will be there. You don’t have to perform maintenance or worry about availability.”

For a business that is built almost entirely on customer trust—after all, no one is going to use a password storage service they can’t rely on to retrieve their stored passwords—finding a dependable partner early was key to 1Password’s success. “We were really, really cautious,” Roustem explains. “People have put a lot of trust in our service and our app. If you screwed up, it would be really hard to recover.” Once the pair knew they could trust AWS, they took advantage of as many of AWS’s offerings as they could, including re:Invent, AWS’s annual learning conference, which Roustem attended for the first time in 2014.

He still beams as he recalls the panel discussions on the topic of, say, migrating to a Virtual Private Cloud. “We used a lot of the stuff that we learned at re:Invent to design the service, to make it as secure as possible, as resilient as possible, to be able to make sure that there is no single point of failure, and that if something goes down, things just heal themselves,” Roustem says.

Today, Roustem has peace of mind knowing not only have they built a secure, resilient service, but it is also robust enough to stay that way as it grows. “We’ve been leveraging Amazon Aurora for a while now and have been really happy with how the databases hold up. We love that as we scale, the service is able to upgrade instance sizes to support more customers.” Even with the millions of existing users who login to 1Password every day, Roustem is confident “tomorrow we could have 10 times more customers, and we still have room to grow. And that feels really nice.”

The 1Password Founders

The company itself continues to grow in the meantime. What started as a two-person team subsisting on ramen noodles numbers almost 500 members today. “Looking back, we didn’t have an HR department, we didn’t have a finance department, and we didn’t have a sales team. There are so many things we didn’t have,” Roustem recalls. “We had to build them from scratch.”

It’s exciting, but for someone who was used to doing it all himself, Roustem is now faced with a new challenge: figuring out how to best manage the growing team. “We have so many smart and talented people, and I’m just trying to stay out of the way these days,” he laughs. “That’s my job.”

Putting Your MVP on the Path to Success (Webinar)

AWS Editorial Team — Fri, 05 Nov 2021 17:34:25 +0000

This post originally appeared on VentureBeat.

The biggest challenge for new startups is to get to market as quickly as possible before the money runs out — and prove that your company is set to take over the market. To accelerate time to market, you need a minimum viable product (MVP), or the smallest and quickest functional version of your idea. It’s something that can be tested quickly and easily, shipped and iterated as often as necessary, as feedback comes back. It proves not only that your new idea works, but that you have the ability to produce it, and the proof that your audience wants it.

The concept comes from Eric Ries, author of The Lean Startup. He calls the MVP “that version of a new product which allows a team to collect the maximum amount of validated learning about customers with the least effort.”

Right from launch, a startup has to jump into the build-measure-iterate cycle, and the companies that can do that quickly and efficiently are the ones that dominate the market. That requires the kind of solid cloud tools and services that let web and mobile developers build scalable, full-stack applications — like AWS, which has a large, versatile suite of services, from what you need to launch your project to helping you scale as you gain market share.

It offers powerful support for agile project development with tools like AWS CodePipeline, which automates code deployment and fosters incremental development, and analytics tools like Amazon Kinesis to keep track of milestones. To continue to grow and deliver on customer feedback fast, there are tools like AWS CodeBuild, which can process and test multiple code builds at once, or AWS CodeStar, which allows you to quickly build and deploy new code.

AWS Amplify offers open-source client libraries. These provide use-case centric, declarative, and easy-to-use interfaces across different categories of cloud-powered operations so mobile and web developers can easily interact with serverless cloud backends, which are created using a visual Admin UI or a Command Line Interface (CLI). That abstracts and leverages a core set of AWS cloud services, which give you authentication, APIs, offline data, analytics, push notifications, and bots at high scale. Amplify lets a single individual manage the entire stack for a given feature, from front-end components to a serverless cloud backend.

Amplify also has a large and responsive community of front-end and full stack app developers who offer support and insight, and a collection of in-depth tutorials.

Startups like Duolingo and SmugMug used AWS to get to market much more quickly, and on November 10 at 10:00AM PT in this VB Live event, you can learn the secrets behind the platform. You’ll hear how AWS enables startups to lower costs, become more agile, and innovate faster, learn how to navigate the AWS cloud platform, choose the right technology for your use case, and successfully develop a minimum viable product in order to propel your startup on the path to success.

Plus, by signing up for AWS Activate, you’ll receive free tools, resources, architecture guidance, technical support, and info on how to access up to $100,000 in AWS credits. Want to get started on AWS Activate right now? Head over here.

For startups, building a minimum viable product can be overwhelming, but the AWS cloud platform helps lower costs, improve agility, and speed up innovation. In this VB Live event, learn how to successfully develop an MVP with AWS and propel your company on the path to success.

Reserve your spot here for free.

On November 10, you’ll learn:

Best practices for developing a minimum viable product
How cloud services make product development more agile
How to increase functionality without rearchitecting, and get to market fast
How to take advantage of the AWS Activate program for free tools and credits

Speakers:

Kumar Chivukula, Co-Founder & CTO, Opsera
Jarrett Wendt, CEO, Spoke Safety
Chris Munns, Technical Leader/Advisor, Startups, Amazon Web Services
Trent Dethloff, Senior PMT, AWS Activate, Amazon Web Services
Seth Colaner, Moderator, VentureBeat

What Does It Mean to Be an AWS Admin at a Startup?

AWS Editorial Team — Wed, 03 Nov 2021 20:40:37 +0000

Guest Post by Faisal Farooq, Startup Solutions Architect and Abhi Singh, Sr. Security Solutions Architect

Most startups work at a high velocities, which can mean that release timelines often overshadow the security foundation. A frequent byproduct of this fast-paced culture can mean that the responsibilities of an Account Admin are not as defined or the role is distributed among several team members. As the team size grows and the company gains momentum, a startup’s customers often require the company to enforce least privilege and clearly define who and what an Admin should do, leaving founders to backtrack architecture-level decisions that were made early on, creating friction and disrupting the business. In this blog post, we will define what an Account Admin’s responsibilities should look like, the training that person, referred to hereafter as an Admin, should have to be effective in their role, and the continuous impact they make on the company.

Overview

An Account Admin is the second most powerful user in the company after the root account. As a result, it’s important to define what that person or entity can do to assign appropriate privileges to the person. Some of the key responsibilities include:

Providing effective utilization of AWS resources and services.
Installing and maintaining your AWS environment by following best practices.
Configuring and maintaining adequate security parameters.
Implementing an information assurance policy, procedures, and reporting.

Admin responsibilities

As mentioned above, the Admin is a trusted advisor for the development teams and founders. Some of the key tenets for a solid Account Admin are:

Provide effective utilization of AWS resources and services

An Account Admin should be the internal trusted advisor and a subject matter expert on AWS. He or she needs to stay up-to-date on recent service announcements and help management realize the most efficient and effective usage of their AWS investment. Day-to-day activities for an Admin include:

Consulting and advising development and product teams on architectural design patterns, and choosing appropriate services with appropriate features to meet business objectives, such as caches, buffering, and replicas.
Building metrics and corresponding alarms using AWS Services, like Amazon CloudWatch to monitor the resource usage and patterns.
Leveraging data from Cost Usage Reports to recommend cost optimization strategies, such as Amazon S3 storage tiers, Amazon EC2 instances, and AWS Lambda resources.
Identifying the opportunities to use AWS Managed Services and Solutions.
Developing deployment strategies to meet business objectives, including resource provisioning, application migration or deployment, and patch management.

Install and maintain the AWS environment by following the AWS Well-Architected Framework

An AWS Admin strives to keep the startup’s environment closely aligned to AWS best practices relative to the Services consumed and Well-Architected Framework, paying close attention to the following:

Architect an automated, cost-effective back-up solution that supports business continuity across multiple AWS Regions.
Determine an architecture that provides application and infrastructure availability and recoverability in the event of a service disruption or failure.

Configures and maintains adequate security parameters

Security is a top priority for the Account Admin. As a result, they are responsible for performing the following key activities to maintain the startup’s environment to the required security standard:

Establish user and resource baselines.
Establish security boundaries between the organization’s resources and partners, customers, and internet.
Evaluate the organization’s AWS environment for security and configuration vulnerabilities.
Based on the organization’s security and compliance requirement, deploy appropriate security controls for public facing web applications, enforcing least privileges across AWS accounts, users, and applications, managing credentials securely. Refer to the Security Pillar for reference.
Deploy automated solutions to notify and correct the baseline deviations, including anomalous access to resources and defining data flow patterns.

The AWS Security Pillar provides detailed guidance on how to configure an Account. Consider testing your configuration using the Well-Architected labs in a sandbox environment for more hands-on training.

Implements the information assurance policy, procedures, and reporting

External regulations and standards may apply to the startup, along with internal ones. The Admin is responsible for identifying appropriate mechanisms, such as AWS Audit Manager, AWS Config, or AWS Security Hub to enforce the security and compliance reporting requirements such as NIST 800-83, HIPAA, FedRamp, PCI etc. to enable continuous compliance. They may also automate the compliance reporting to appropriate parties. AWS provides several services and solutions that can enable the Admin achieve the compliance and assurance goals. Based on the needs of the startup, the Admin can leverage these turn-key solutions and guides that align with AWS’s guidance.

Continuously improves the AWS environment for efficiency, effectiveness, security, and cost

Depending on the level of understanding of AWS and the startup’s business, the Admin continuously reviews the existing environment and identifies improvement opportunities. They should review the performance metrics against the target and suggest improvements, identify the bottlenecks and recommend alternate strategies. They may also test the automated deployment and rollback strategies, and leverage and promote infrastructure as code and CI/CD pipeline-based deployment. The Admin will regularly review security related issues and develop automated mitigation mechanisms such as alternate architectures for repetitive items, and lastly, maintain Runbooks and Playbooks for common tasks highlighted above. The AWS Well-Architected Framework has an Operational Excellence pillar that can help Admin’s with continuous improvement

How can you innovate in your role as an AWS Admin?

AWS routinely provides guidance and architecture patterns, based on trends across the industry. This guidance along with the items listed below will help Admins invent and simplify on behalf of your customers:

Review the AWS latest guidance via re:Invent, whitepapers, and Tech Talks to identify new patterns.
Automate day-to day-tasks using AWS SDKs, services like SSM Documents, and SSM Automation.
Coach application teams on newly released services and design and architecture patterns.
Reduce costs by working with your AWS Account teams on leveraging different mechanisms such as Reserved Instances, Saving Plans, service-specific volume discounts like the CloudFront security savings bundle, and Enterprise Discount Programs.
Automate a Business Continuity and Disaster Recovery strategy.
Develop automation for site reliability against defined targets.
Conduct periodic Well-Architected Reviews on your workloads.

Conclusion

Startup management often grapple with defining and outlining an AWS Admin’s responsibilities. With these resources, founders and stakeholders can hire and develop the best, while maintaining an appropriate level of separation of duties that often plague a startup’s growth and scalability. By following the practices defined above, AWS Admins can balance performing day-to-day activities effectively, help their startups adopt good security hygiene practices often required as part of third-party assurance, and optimize infrastructure costs.

	Faisal Farooq is a Startup Solutions Architect who focuses on the Security and Compliance. He routinely hosts customer open forums to help Startups to discuss the industry wide challenges. He is passionate about helping customers use AWS more efficiently, effectively, and securely.
	Abhi Singh is a Senior Solution Architect who specializes in security and compliance within AWS. He has over 20 years of experience in information technology consulting and leadership experience.

Serverless Tenant Isolation in SaaS Applications with SigTech

AWS Editorial Team — Fri, 29 Oct 2021 18:21:06 +0000

Providing innovative technology solutions for some of the world’s leading investors, security is at the heart of everything SigTech does. SigTech offers future-proof quant technologies to global investors. Cloud-hosted and Python-based, the platform integrates a next-gen backtest engine and analytics with curated datasets covering equities, rates, FX, commodities, and volatility. SigTech eliminates the expensive upfront costs of infrastructure build-out, giving clients an edge in alpha generation from day one.

Combining defense-in-depth with automation allows SigTech to meet their security obligations while developing at pace. Through the use of the deep and rich features provided by AWS services, SigTech has been able to build quant technologies that are operable by their engineering team, can be developed and iterated in an agile way, and meet the security requirements of their customers. Their aim is to make the research process as efficient as possible by solving common challenges encountered by systematic traders and researchers.

SigTech’s Solution

Our platform allows systematic investors, such as hedge funds and asset managers, to research, implement, and deploy cross-asset strategies on a single end-to-end platform. Once on-boarded, customers are able to provision environments, known as workspaces, to research and develop their trading strategies. Each workspace allows customers to rapidly provision preconfigured Amazon SageMaker Jupyter notebook research environments with standard and customized datasets immediately available. Quant researchers and traders can then develop and test their strategies within these research workspaces before deploying their models.

When building the SigTech SaaS quant platform, we had several key challenges to consider. The first was to align our platform development methodology with the business focus on rapid time to market. This meant ensuring that the efforts of our team were focused on the development of new, high-quality features, rather than operating manual processes. We therefore took the decision to automate everything possible, including key customer activities such as the on-boarding and provisioning processes.

Operating within the highly-regulated financial services industry, security of our data and our customers’ data and intellectual property is a top priority, so layered security controls have been implemented throughout the platform to meet our customers’ expectations around the protection of their data and trading algorithms.

In order to operate these controls while enabling high quality agile feature delivery, we had to ensure our Security Operations (SecOps) capability had full visibility of our platform and could quickly respond to any risk or incident identified. We decided to gather observability data at scale, and its deep monitoring became a key priority, where leveraging automation in data collection, monitoring, and alerting systems for SecOps became essential requirements.

The need to leverage the Cloud to enable a highly scalable and secure platform that could be continually developed to a high standard in a fast-paced environment became self-evident. We chose AWS for its breadth and depth of services, API and automation-first strategy, and strong security controls.

Product Design

One of the primary considerations for SigTech as a SaaS platform is data security and ensuring that customer data is completely segregated. Each user is given a Research Environment which is based on JupyterLab and powered by Amazon SageMaker. This environment comes preconfigured with the SigTech Quant Framework and access to our data lake containing more than 600,000 instruments and datasets, covering every major asset class and alternative dataset from the world’s leading data providers. Within the Research Environment, users can share workspaces with colleagues on the same tenant and develop and test their strategies before deploying their models to a Production Environment.

We operate our platform in a multi-tenant model, meaning all customer resources exist within the same AWS account and network. Tenant isolation is a fundamental security principle for us, ensuring our customers’ intellectual property is protected. Isolation is achieved by providing each of our customers with unique, auto-generated AWS Identity and Access Management (IAM) roles, and private and unique encryption keys by using Customer-Managed Keys (CMKs) – within the AWS Key Management Service (KMS) to encrypt/decrypt their data. The data is stored in unique Amazon Simple Storage Service (Amazon S3) buckets for each tenant with S3 bucket policies restricting access to only that tenant’s resources.

The Jupyter notebook environment is provided by Amazon Sagemaker Notebook Instances for each tenant. Inbound and outbound network controls are configured in the form of Security Groups. These groups ensure that the instances are isolated from other tenants and prevent any data leakage outside of the environment by blocking all external communication.

All of the platform logs are strictly monitored in a centralized logging location using Amazon S3 and Amazon CloudWatch to store the AWS service’s logs and AWS CloudTrail for monitoring all user activity on the AWS APIs. This is combined with Amazon CloudWatch Alerts & Amazon QuickSight dashboards to provide proactive monitoring.

Solution Walkthrough

Three elements made the SigTech platform possible and allowed us to address the key design challenges:

Automated customer tenant on-boarding
Multi-layered security controls, and
Monitoring and observability at scale.

Figure 1. SigTech’s overall solution design

Automated Tenant Onboarding

Given the requirement to create specific AWS resources for each customer, develop at pace, and ensure security through all aspects of our platform, it was necessary that the process of on-boarding new customers or tenants was repeatable, auditable, and automated. The capacity to scale efficiently, irrespective of the number of customers, was also necessary.

The process would require several steps and multiple calls to APIs to update tenant resources, as the process itself generated new resources. For example, setting up tenant IAM roles, CMK keys in KMS, and S3 buckets would require the additional updating of policies to restrict tenant access to only these resources. Therefore, coordinating these steps in an automated fashion was critical. We identified the stateful workflow management of AWS Step Functions as being the right fit to manage these processes. It proved invaluable in automating our tenant on-boarding process end-to-end without requiring any manual steps.

Figure 2. Business workflow when onboarding a new tenant

We developed AWS Lambda functions that execute various steps in the Step Functions workflow. This allows the steps to be run multiple times until the final desired state is achieved, with the current state of workflow execution and any required retries automatically managed by the Step Functions execution logic.

Purpose of Workflow

In order to achieve a secure multi-tenant environment, SigTech uses a range of AWS services to provision resources for a specific customer and then utilizes the power of IAM roles & permissions to ensure that they are the only tenant who can read that data.

As part of the new tenant workflow, each customer is given their own IAM role & KMS key, as well as an S3 bucket to store all of their custom data. The customer’s IAM role ensures that whether starting up a research environment (powered by SageMaker) or executing a strategy (using AWS Batch & AWS Fargate), security is maintained.

Figure 3. Step Function workflow and associated AWS services used to configure a new tenant

When a new tenant is configured in the SigTech Platform administration portal, a new workflow is triggered to set up the various resources required for that tenant. During the workflow, we persist any state required (e.g., ARN details for newly created resources) in the event that this can be utilized in any subsequent downstream workflow steps. Once everything has been successfully configured, then the final step in the workflow persists any resource ARNs that we need to reference in the future in our tenant Amazon DynamoDB table.

Retrying Failed Workflows

One issue that we encountered when running the tenant on-boarding Step Function workflow was eventual consistency between the IAM cache in some AWS services. Attempts to reference an IAM role or policy, newly created in the previous step, would fail because the new role/policy reference had not yet been replicated for use with other services. The solution here was to simply retry the failed step a few seconds later, including a back off rate to ensure eventual consistency. This required no code change to our existing Lambda logic, but instead was just a Step Function configuration change.

Should a Step Function continue to fail, this would suggest a more complicated underlying issue. In this instance, Step Functions are an ideal architecture, as we can alert the SigTech team to the problem, displaying the logs and fixing the underlying data to resolve the issue or escalate it to another team.

Multi-layered Security Controls

As mentioned in the Solution Design section, layered security design is used throughout the platform to ensure tenant isolation and data protection. This approach leverages dedicated authorization, encryption, and data storage resources for each tenant, networking controls for workload isolation, and deep monitoring capabilities to provide visibility of our controls in practice. It is necessary to consider each of these in more detail.

Identity and Access Management (IAM) roles

During the automated tenant on-boarding process, we create unique Identity and Access Management (IAM) roles for every customer, each with IAM policies providing access only to resources belonging to their tenant. This ensures our customers’ users are protected and isolated from other customer entities and can only access and manage their own SigTech environment and data. The IAM policies configured for these roles have “Allow” actions restricted to tenant-specific resources by explicitly referencing the Amazon Resource Name (ARN) within the policy. For example, the IAM policy snippet below shows how a customer’s IAM role policy gives them access to only manage their assigned S3 bucket.

{
            "Sid": “List customer S3 bucket”,
            "Effect": "Allow",
            "Action": "s3:ListBucket",
            "Resource": "arn:aws:s3:::sigtech-12345678-abcd-1234-1234-1234567890abc"
        },
        {
            "Sid": “Manage Objects inside the customer S3 bucket”,
            "Effect": "Allow",
            "Action": [
                "s3:PutObject",
                "s3:ListMultipartUploadParts",
                "s3:GetObject",
                "s3:DeleteObject",
                "s3:AbortMultipartUpload"
            ],
            "Resource": "arn:aws:s3:::sigtech-12345678-abcd-1234-1234-1234567890abc/*"
}

Figure 4. IAM policy enabled to manage only customer assigned S3 bucket

Furthermore, each IAM role has a trust policy configured to enable administrative service entities to assume the tenant specific role and perform actions on services like Lambda and SageMaker. The next block of code demonstrates an example of this functionality.

   {
      "Sid": "Assume Role Permissions",
      "Effect": "Allow",
      "Principal": {
        "Service": [
          "lambda.amazonaws.com",
          "sagemaker.amazonaws.com"
        ],
        "AWS": [
          "arn:aws:iam::<account_number>:role/<name_of_role_1>”,
          "arn:aws:iam::<account_number>:role/<name_of_role_2>"
        ]
      },
      "Action": "sts:AssumeRole"
    }

Figure 5. IAM trust policy configured to give permissions to other IAM roles

Encryption Key Controls

We also further isolate customer data through encryption key controls. All customer data on our platform is encrypted at rest using dedicated per-tenant Customer-Managed Keys (CMKs) provided by the AWS Key Management Service (KMS). This provides private and unique encryption keys to encrypt/decrypt customer’s data. Each tenant is assigned a CMK with a key policy configured to restrict usage of the key to tenant-specific IAM roles. The following example shows how the KMS policy allows only the customer-specific IAM role to use the key.

{
            "Sid": "Enable SigTech IAM User ",
            "Effect": "Allow",
            "Principal": {
                "AWS": "arn:aws:iam::<account-id>:role
/sigtech-12345678-abcd-1234-1234-1234567890abc”
            },
            "Action": [
                "kms:Encrypt",
                "kms:Decrypt",
                "kms:ReEncrypt*",
                "kms:GenerateDataKey*",
                "kms:DescribeKey",
                "kms:CreateGrant",
                "kms:ListGrants",
                "kms:RevokeGrant"
            ],
            "Resource": "*"
}

Figure 6. KMS key policy enables a specific IAM role to use the key

The IAM role policy is configured to ensure it has the right permissions to manage only the tenant specific KMS key. An example of this configuration can be observed in the following code block.

{
            "Sid": "Enable SigTech KMS Key",
            "Effect": "Allow",
            "Action": [
                "kms:GenerateDataKey",
                "kms:Encrypt",
                "kms:Decrypt"
            ],                  
     	    "Resource":"arn:aws:kms:<region>:<account-id>:key
                     /12345678-abcd-1234-1234-1234567890abc”
}

Figure 7. IAM policy enables KMS key permissions.

These controls ensure that tenants can only access their own encryption key, and the encryption key can only be accessed by that tenant. The data encrypted with these keys can then only be decrypted with the tenant’s authorized users with access to those IAM roles.

Amazon SageMaker

In order to ensure that a customer’s Research Environment is only accessible by their users, we use an IAM role specific to that tenant to run the SageMaker instance. This is achieved using SageMaker instance tags.

   {
         Sid: "Sagemaker",
         Effect: "Allow",
         Action: [
           "sagemaker:ListTags",
           "sagemaker:CreatePresignedNotebookInstanceUrl",
           "sagemaker:StopNotebookInstance",
           "sagemaker:DescribeNotebookInstance",
         ],
         Resource: `arn:aws:sagemaker:*:${awsContext.account}:notebook-instance/*`,
         Condition: {
           StringEquals: {
             "sagemaker:ResourceTag/TenantId": “sigtech-12345678-abcd-1234-1234-1234567890abc”,
           },
         },
       },

Figure 8. IAM policy restricting tenants to their own SageMaker instances

Data Storage & Access Controls

Alongside authorization and encryption controls, controls at the data storage layer provide an additional security barrier. Each customer is issued with a dedicated S3 storage bucket to store their data. The bucket is configured to automatically use the customer’s dedicated CMK to perform server-side encryption of all data stored in the bucket. The S3 bucket uses bucket policy entries to restrict access to only the tenant’s resources.

{
            "Sid": "SigTech IAM Role access",
            "Effect": "Allow",
            "Principal": {
                "AWS": "arn:aws:iam::<account-id>:role/org/
                             sigtech-12345678-abcd-1234-1234-1234567890abc”
            },
            "Action": [
                "s3:AbortMultipartUpload",
                "s3:DeleteObject",
                "s3:GetObject",
                "s3:GetObjectAcl",
                "s3:GetObjectTagging",
                "s3:GetObjectVersion",
                "s3:GetObjectVersionAcl",
                "s3:GetObjectVersionTagging",
                "s3:ListMultipartUploadParts",
                "s3:PutObject",
                "s3:PutObjectAcl",
                "s3:PutObjectTagging",
                "s3:PutObjectVersionTagging",
                "s3:RestoreObject"
            ],
            "Resource": "arn:aws:s3:::sigtech-12345678-abcd-1234-1234-1234567890abc/*"
        }

Figure 9. S3 Bucket policy enables specific IAM role to manage the bucket

These bucket policy controls ensure that only specific tenant’s resources can access that bucket, and the bucket only allows access from resources owned by that tenant. For example, when a customer runs a research environment via the SigTech platform, which launches a SageMaker Notebook Instance, that tenant’s IAM roles are attached to that compute environment. This, in turn, restricts access to that tenant’s CMK and authorizes access to that tenant’s S3 storage bucket only. The tenant can then access and decrypt the data, but only from that S3 bucket and only via that CMK.

Below is an example flow diagram showing the IAM policies, CMK policies and S3 bucket policies restricting access between different tenant resources.

Figure 10. Permissions flow preventing access between resources.

Network Controls

The SigTech platform makes use of the Amazon SageMaker Notebook Instances, which run on top of Amazon EC2 Virtual Machines. To ensure the protection of our customers’ data and specific algorithms, we employ network controls at the EC2 and VPC level to ensure strict data isolation. To restrict data exfiltration, we use dedicated EC2 Security Groups to isolate the tenant SageMaker Notebook instances in a virtual firewall where we can control inbound and outbound traffic. We control the access to each S3 bucket on the network layer using the VPC Endpoint policy feature (presented below), where we allow access to the S3 buckets only from inside the SigTech AWS environment.

{
  "Statement": [
    {
      "Action": "*",
      "Effect": "Allow",
      "Resource": "arn:aws:s3:::*",
      "Principal": "*",
      "Condition": {
        "StringEquals": {
          "aws:PrincipalOrgID": [
            "<org_id>"
          ]
        }
      }
    }
  ]
}

Figure 11. VPC Endpoint policy restricting access to S3 buckets

These tenant-specific isolation controls are summarized in the figure below, which describes the separation of customer data and the differing CMK permissions between a customer role and a SigTech role.

Figure 12. Layered Security Controls on the SigTech platform

Monitoring Controls

To ensure that our tenant isolation and data protection controls are operating as expected, the SigTech platform is also monitored and protected by various AWS security services. These include:

AWS WAF and AWS Shield (for platform-level network firewall and DDoS protections)
AWS GuardDuty (for networking monitoring and threat detection)
AWS Security Hub (for security log and event collection)
Amazon Inspector (for vulnerability assessment)

These services all feed log and insight data to a separate, dedicated Security AWS account for auditing, monitoring and isolation from customer and platform environments. This ensures that security data cannot be tampered with by non-security-focused individuals. All the logs from these services are safely stored in a Centralized Logging S3 bucket in this Security account and used for monitoring and alerting processes, see below.

Figure 13. SigTech’s extra layered Security Controls

Monitoring and Observability

For monitoring and audit purposes, each AWS Account stores the logs of several different services in an S3 Bucket. These services include Amazon API Gateway, Amazon CloudFront, Amazon Route53, Amazon Virtual Private Cloud (Amazon VPC), and AWS WAF. To enable the SecOps team to effectively manage our environments, all of the logs are replicated in the dedicated Security AWS account. There, they can be combined and searched using a variety of AWS tools.

An Amazon Athena database is used for all the “Logs” S3 buckets in the Security account, where data can be queried efficiently using SQL queries. This enables the SecOps team to monitor metrics such as rejected SSH connections or requests blocked by the WAF for different purposes, such as someone attempting to exfiltrate data from the platform.

Figure 14. SigTech Centralized Logging System

To make the workflow efficient for SecOps operators, we use QuickSight dashboards. These leverage the Athena databases to visualize the data and empower our SecOps team to gain deeper insights into the data and to identify signals. The code block and dashboards below provide an example of a query used by the SecOps to find blocked HTTP requests within Athena and visualize these results in QuickSight.

SELECT <waf-logs-table>.*, <waf-names-table>.*
FROM 
    (SELECT httprequest.clientip AS clientIP,
         webaclid AS webAclId,
         httprequest.country AS country,
         httpsourceid AS dest,
         from_unixtime(timestamp / 1000) AS time,
         COUNT(*) AS count
    FROM "<database-name>"."<waf-logs-table>"
    WHERE action='BLOCK'
            AND timestamp >= to_unixtime(current_timestamp - interval '7' day)
    GROUP BY  httprequest.clientip, httprequest.country, httpsourceid, webaclid, from_unixtime(timestamp / 1000)
    ORDER BY  count desc) AS waf_logs
LEFT JOIN 
    (SELECT name,
         webaclid AS id
    FROM "<database-name>"."<waf-names-table>") AS waf_names
    ON webAclId=id

Figure 15. QuickSight dataset created with Athena SQL Query

Figure 16. QuickSight dashboard showing Rejected SSH requests

Figure 17. QuickSight dashboard showing users blocked by SigTech WAF rules

Conclusion

Deploying a secure platform is not a one-off exercise; SigTech continues to focus on enhancing our isolation and monitoring controls as part of our day-to-day operations and development activity. We are proud to work closely with our AWS account team, AWS Solutions Architecture teams, and AWS service teams to provide feedback and keep up to date with the latest releases and capabilities.

Further resources you may find useful on the topic of building SaaS platform construction on AWS and multi-tenant security include the AWS SaaS Factory Bootcamp, this blog on implementing SaaS tenant isolation with ABAC and AWS IAM, SaaS Journey Framework whitepaper, and AWS SaaS Boost.

	Adam Temple is a Senior Solutions Architect at AWS, working with mid-size enterprises across a variety of industries in the UK. He has over 8 years of experience helping customers transform their businesses through cloud adoption with a focus on DevOps adoption, Data & Analytics Modernization and AI & Machine Learning. When not working, he enjoys Formula 1 and learning new things whenever possible.
	Nitin Tiwari is a Senior Solutions Architect at AWS, working with customers of all sizes across the UK, helping them across the breadth of the AWS Services. He has an area of focus in the topics of Security & Compliance and Financial Services. When not working with customers or thought leadership projects, he likes to spend time with family and enjoys sports and travel.
	Brenda Szamosi is a Security Engineer at SigTech, responsible for maintaining the security of the platform by developing and monitoring systems to detect and prevent any potential security breaches. Outside of the office, she likes to spend time with her family, travel to new places or practice paddleboarding.
	Tim Glass is Head of Platform at SigTech and enjoys developing new and innovative technical solutions to solve their customer’s problems. When not at work he can be found going on adventures with his two young children, or watching YouTube videos to try and learn how to renovate his home.

Applications Are Now Open in the UK for the AWS Healthcare Accelerator Programme

AWS Editorial Team — Wed, 27 Oct 2021 21:34:26 +0000

By Dr. Rowland Illing DM MRCS FRCR, Director & Chief Medical Officer, International Public Sector Health, AWS

The United Kingdom (UK) healthcare system’s acceleration of innovation in service delivery, along with its increased adoption of digitally-enabled, secure, and compliant solutions through the Covid-19 crisis has brought about great benefit for patients. Embracing digital health innovations has enabled the NHS to deliver a world-leading vaccination programme showing their inherent talent and collaborative, patient-centric capacity to transform at pace. Now, their focus moves to tackling backlogs in elective care, continuing to implement the NHS Long Term Plan, and focusing on transformation of services to support NHS resilience. At the same time, the health and care system is seeking to optimise the proliferating volume of healthcare data to further improve patient outcomes and operational decisions.

To support both high-potential healthcare startups and the UK healthcare system’s demand for these types of solutions, we are excited to announce the launch of the AWS Healthcare Accelerator programme in the UK.

AWS Healthcare Accelerator

The AWS Healthcare Accelerator is a four-week technical, business, and mentorship accelerator opportunity, open to UK-based healthcare startups or international healthcare startups that have existing UK operations. AWS will deliver the AWS Healthcare Accelerator programme in the UK in collaboration with PUBLIC, a leading government-focused technology company, who bring experience in delivering accelerator programmes, a strong network of healthcare startups, and an international presence.

The cohort of 10 healthcare startups with an established product-market fit and existing customers and revenue, who seek to use AWS to help solve the biggest challenges in the healthcare industry, will be selected to participate in this new programme. It is tailored to accelerate growth in the cloud, focusing on innovation in the areas listed below, such as remote patient monitoring, mental health support solutions, care navigation, analytics supporting triage and prioritisation, and workforce productivity tools. Applications open today and must be submitted no later than December 5th.

The goal of the AWS Healthcare Accelerator is to cultivate and promote innovative startup solutions that achieve the Quadruple Aim of improved patient experience, improved clinician experience, better health outcomes, and lower cost of care. In seeking out secure, compliant, data-driven healthcare solutions, AWS will support public sector healthcare enterprises to accelerate their digital transformation.

The AWS Healthcare Accelerator will be open to healthcare startups targeting the following areas:

Citizen Access
Innovative technology offerings to improve citizen access to health and social care services, including digital front door solutions that streamline how citizens access health and social care services, and simplify processes to navigate services across care settings. We are looking for truly citizen-centric innovations, particularly those that promote equity of access.

Elective Care Recovery
Startup solutions to help the NHS with elective care, cancer care services, and management of the increasing demand on mental health services. This includes waitlist management, triage/prioritisation, clinical decision support tools, as well as innovation supporting the diversification of care delivery models outside the hospital setting.

Population Health and Precision Medicine
Digital health products seeking to support the NHS segment, with focus on the prioritisation and delivery of population health and care management initiatives. This includes innovations across the population health value chain, from planning (e.g. population health analytics) to delivery (e.g. technology-enabled clinical intervention programmes). We are also interested in hearing from startups working in clinical genomics research developing tools to provide person-centric care.

Workforce
Innovation aimed at promoting new ways of working that make use of the full range of people’s skills and support the care workforce to improve productivity, efficiency, and effectiveness.

AWS and PUBLIC will select 10 startups to participate in the four-week accelerator that will take place virtually. Selected startups will receive AWS Promotional Credits, specialized AWS training, mentoring from healthcare domain and technical subject matter experts, business development, go-to-market guidance, and investment guidance, and potential proof of concept opportunities with public sector healthcare customers. The programme will bring in healthcare industry leaders to collaborate with startups on topics ranging from defining business models, regulatory pathways, clinical validation, Electronic Health Record integration, and more. There will also be collaboration opportunities with AWS customers and members of the AWS Partner Network looking for innovative healthcare solutions.

Applications will be judged on several factors, including the innovative and unique nature of the project, the overall value the solution will bring to the healthcare industry, the creative application of AWS to solve problems, and the team’s ability to deliver on an identified opportunity. The evidence shows many groups are under-represented in founding and scaling health technology companies. At AWS, we welcome applications from founders of all genders, ethnicities, and backgrounds, who in turn reflect the diversity of people within and served by the UK’s Health and Care system.

Learn more about how to apply to the AWS Healthcare Accelerator and how AWS for Health supports customers and partners. And, see how PUBLIC helps enable healthcare startups to scale and succeed.

Build a Hybrid Architecture for Local Compliance and Global Scalability

AWS Editorial Team — Mon, 25 Oct 2021 19:32:49 +0000

By Saud Albazei, Alexandru Costescu, and Anshuman Nanda

Startups at any stage face many regulatory challenges when expanding to new markets or trying to comply with data residency regulations in their home market, putting them at a disadvantage compared to established enterprises that can afford the upfront capital investment of operating their own physical facilities with AWS Outposts where an AWS region is not yet available in a country. In this post, we will explore alternatives where a startup could run workloads in multiple infrastructures in a hybrid approach to comply with local data residency requirements, while utilizing the AWS regions for global scalability.

Understanding the regulation is fundamental to implementing a compliant technical solution. Different countries impose different regulatory requirements, and some even regulate industries differently. That could include Logistics, FinTech, HeathTech, EdTech, and a few others. The regulated part of the data in most cases is a subset of the entire workload. An example of that would be Personally Identifiable Information (PII), Payment or Financial Transactions, Short Message Services (SMS) and more. Once your local regulations have been clearly identified, you can use a few AWS services to help with the deployment, operation, and monitoring of all infrastructures in a centralized location.

This hybrid architecture gives startups the freedom to provision any infrastructure as long as the servers are running a supported operating system such as Amazon Linux 2, Bottlerocket, Ubuntu, RHEL, SUSE, Debian, CentOS, or Fedora. This flexibility allows you to run the regulated workloads anywhere faster and more cost effectively.

Primary Approaches for Hybrid Architectures

There are two primary scenarios in which this architecture can be implemented. The first use case is when operating in a single country with a single-entry endpoint such as a local API Gateway that routes requests based on the data classification. The major benefit of this approach is that the clients can continue to communicate to the back-end over a single endpoint and API.

The second use case is when operating in multiple countries with multiple entry points and multiple API gateways. In this approach, the client has to have the necessary routing logic to communicate with the services in the user’s respective country. An example to consider would be a FinTech startup that has to store user’s financial transactions based on the country they reside in. Each local deployment will have a dedicated endpoint, and the client will communicate to each local endpoint based on the location set in the user’s profile.

All of these local and Cloud based deployments are managed in a centralized location, and in the following demonstration, you will learn how to deploy a sample local containerized service, focusing on the following tools:

AWS Systems Manager – A unified user interface to track and resolve operational issues across all applications in a central place.
Amazon Elastic Container Service (Amazon ECS) – A fully managed container orchestration service.
Amazon CloudWatch – A monitoring and observability service.

The first step is to provision the local servers with the desired capacity and supported OS. Once you have them up and running, you can launch your first Amazon ECS Anywhere Cluster. ECS Anywhere is a feature of Amazon ECS that enables you to easily run and manage containerized workloads on your managed infrastructure.

Once the ECS cluster is created, you can register local servers to it as ECS Instances, adding the provisioned local servers to this cluster as hosts. The control plane will reside in the Cloud and continues to be fully managed by Amazon ECS. After registration, it will give you a command to execute a shell script that will install all required software and register the instance with AWS Systems Manager and the ECS Cluster. Local servers will host the data plane and an ECS Agent, and the control plane connects securely to the hosts over a TLS connection, and no data is transferred between the control plane and host other than deployment related metrics and health status of the services.

Now, you have an ECS Cluster with registered local instances and ready for a deployment. Containerized deployments in ECS are packaged in tasks, which share many characteristics with Kubernetes pods. You could deploy a single task, or multiple tasks in services distributed across multiple instances for high availability and managed scalability. In order to deploy the local service, you need to create a task definition for it first. You can do that by navigating to Task Definitions and creating a new definition of type external. Fill in all necessary information, and add the necessary containers and their configuration and they could either be registered with Amazon Elastic Container Registery (Amazon ECR) or any other container registry service such as Docker Hub.

Your ECS cluster is now ready for the deployment. Create the service, which will be type ‘external.’ Set the service name, desired number of tasks, deployment type, and placement strategy. Once created, Amazon ECS control plane will work with the deployed ECS agents in each local server to download the container images and run them, which includes checking their health, and reporting their state in the AWS console.

Now, you have your containerized application running on your locally provisioned servers. The application is ready to take in requests from a frontend API Gateway, Reverse Proxy, or a Load Balancer such as NGINX or KrakenD through the exposed ports. Additionally, given the ephemeral nature of these containers, the application will need to persist data such as databases or media objects. Amazon S3 can be used for object storage if this media is not regulated, or use client-side encryption before uploading objects to S3 if the regulation allows. For the database, Amazon RDS on VMware can be used locally, or managed by the customer through deploying it to a dedicated server managed and monitored by AWS Systems Manager and Amazon CloudWatch respectively. Both services can help manage the server fleet, automate operations, and offer a centralized observability solution for all workloads.

Observability for Hybrid Deployments

All registered servers, either with the ECS cluster or with SSM, can be monitored and managed through the AWS console. This includes health status of the servers by observing performance metrics and connecting to them over SSH. The CloudWatch agent also allows you to collect internal system-level metrics from your servers as well as collect logs to be displayed in the CloudWatch Dashboard. This dashboard is highly configurable, so you can choose what data is uploaded to the Cloud to avoid uploading regulated data. In addition, CloudWatch metrics can trigger Lambda functions that could be designed to respond to certain events such as failover, scaling the local infrastructure and various other events.

Conclusion

This hybrid approach allows you to create the necessary services in a combination of AWS managed infrastructure as well as customer managed infrastructure while maintaining a single observantly location and using the same AWS services and tools to deploy and manage the workloads. Segregating regulated microservices from unregulated ones is key to having a successful hybrid architecture that can be scalable and highly available. Startups can expand quickly and cost effectively using this approach while continuing to benefit from the continuously growing global AWS infrastructure.

	Saud Albazei is a Solutions Architect at Amazon Web Services. He advises startups from early stages through helping them go public. He leverages his experience to help startups innovate and bring ideas to life. He has a passion for building distributed and scalable systems using serverless technologies.
	Alexandru Costescu is a Startup Solutions Architect based out of Bucharest ,Romania. He is passionate about all things tech and about designing systems that can take full advantage of the cloud and embrace the DevOps culture.
	Anshuman Nanda is a Sr. Startup Solutions Architect based out of Dubai, UAE. He enjoys learning new technologies and helping customers solve complex technical problems by providing solutions using AWS products and services.

The AWS Attendee Guide for Startups at Re:Invent 2021

AWS Editorial Team — Thu, 21 Oct 2021 21:04:38 +0000

AWS re:Invent 2021 is back, live and in-person once more. While last year’s three-week digital event was a great way to connect in an unusual time, you can’t beat the atmosphere, energy, and excitement that comes with reuniting in person. We’re not wasting any time jumping back in, either. This year marks a decade of re:Invent, and it’s a truly special milestone. re:Invent 2021 is chock-full of activities, including insightful breakout sessions and interactive workshops, designed to immerse you in the rapidly evolving world of startup culture.

This year, we partnered with AWS Community Hero, Martin Buberl, to create an attendee guide specifically for startup attendees. Martin loves the fast pace of constantly iterating, building, and shipping products that customers love—it keeps him coming back for more. If you love to explore how AWS has changed the game for how rapidly you can turn ideas into products that scale, then his guide is for you.

Here are the highlights of what to expect this year.

Startup Central in Expo

If you’re looking to connect with other startups or dive deeper on a specific challenge, be sure to swing by Startup Central—the startup hub in the Venetian Expo. Hear from some of the world’s fastest-growing startups in Startup Central Theater or drop in to the Ask an Expert Bar to consult with AWS Startup Solutions Architects one-on-one to get your questions answered. Want to learn more about a newly launched feature? Request a personalized walk-through to see it in action at one of the demo stations.

Startup Central Theater Sessions

Session ID	Title	Featuring
STP214	An inside look at Drata’s automated security and compliance	Drata
STP222	How Sofar Ocean is building ocean intelligence at scale	Sofar Ocean
STP219	DayTwo redefines microbiome analysis using AWS frameworks	DayTwo
STP221	Re-inventing biomanufacturing to bring curative therapies to market at scale	Resilience
STP215	Building a serverless banking-as-a-service platform on AWS	Solarisbank
STP220	How fintech startup Flutterwave achieved “unicorn status”	Flutterwave
STP216	Building better: The benefits of a well-architected startup	AWS
STP217	Cost Optimization: Stretching your runway to go further	AWS
STP218	Scaling your B2B startup	AWS

Breakout Sessions

While many re:Invent attendees grab a prime seat for keynotes and leadership sessions, there’s nothing like hearing from someone who’s been where you are now, who put the pieces together and made something big. Breakout sessions give you a chance not only to hear success stories from founders and startup executives, but also to glean insights, advice, and direction for your own journey.

There are countless breakout sessions to explore at re:Invent 2021, and this year, we have leaders from BlockFi, Cambridge Mobile Telematics, Databricks, Drift, ION Energy, Olive, and Starburst Data speaking in the Startup Track. They’ll take the stage to share mechanisms for rapid innovation, preparing for scale, and how they’ve achieved success in a volatile or challenging market.

Session ID	Title	Featuring
STP213	How startups are building solutions for a sustainable future	ION Energy
STP212	Improving the human experience in healthcare with Olive	Olive
STP207	Using privacy-first data processing to make drivers smarter	Cambridge Mobile Telematics
STP205	How running on AWS is paying off for Starburst Data	Starburst Data
STP201	How BlockFi scales to meet demand for crypto-asset services	BlockFi
STP204	Engineering for IPO with scalable architectures	Databricks
STP206	Drift on AWS: Transforming how businesses buy from businesses	Drift

The full re:Invent session catalog is now live and reserved seating for in-person attendees is now available. To get access, register here, or log in to start building your schedule.

Pro-tip: Use the filters find tracks, topics (hint: startups), and services you can’t miss—or simply search for the session ID below.

Interactive Sessions

If you’re attending re:Invent, there’s a good chance you’re someone who likes to ask questions. Don’t worry: we love answering them. It’s why this year’s event has a full agenda of interactive sessions to attend from hands-on guided workshops to chalk talks with AWS Solutions Architects and technical pros. Learn how to work backwards using Amazon’s proven model for launching new products and services. Get the scoop on decision-making at-scale and how to build a team of decision-makers as you grow, or how to successfully exit a startup.

Session ID	Session Type	Title
STP202	Chalk Talk	How leading startups prevent outages, incidents, and escalations
STP212	Chalk Talk	Success secrets for mergers and acquisitions
STP207	Chalk Talk	How startups improve productivity and reduce technical debt
STP205	Chalk Talk	Exploring multi-Region application architecture patterns
STP201	Chalk Talk	Designing the right data strategy for your startup (featuring Dremio)
STP204	Chalk Talk	Automating and evidencing key compliance security controls
STP301	Workshop	Startup security: Techniques to stay secure while building quickly

Join Us at re:Invent 2021

Over the years, re:Invent has played host to some of the industry’s foremost experts and authorities in startup success. And, the event has been a springboard for some of today’s most well-known startups. In addition, startups are using AWS to innovate quickly as they develop solutions to help organizations across all sectors achieve sustainability goals. We’re looking forward to seeing you in-person again at re:Invent 2021. The health and safety of our customers, partners, and employees remains our top priority. Before booking your travel, review the health and safety measures we’ve added this year.

Can’t attend in person? If you can’t make it to Las Vegas this year, register to virtually attend live keynotes and leadership sessions. Plus, we’ll have breakout sessions available on-demand during the live event. Register today to be a part of the action in-person or online.

Share Your Integration Innovations in the Flex Your Skills Contest on AWS

AWS Editorial Team — Wed, 20 Oct 2021 16:34:53 +0000

Think you have a great idea other startups should be using? The Flex Your Skills Contest is your chance to highlight your creativity and ingenuity in integrations in AWS using Step Functions. AWS Step Functions is a low-code visual workflow service used to orchestrate AWS services, automate business processes, and build serverless applications. With the release of AWS SDK Integration in Step Functions, supported integrations for AWS Services in Step Functions have increased from 17 to over 200, and supported AWS API Actions have increased from 46 to over 9,000 making it even simpler to build on AWS.

Innovation and experimentation are core functions for Startups, and we believe the launch of AWS SDK Integration creates exciting new possibilities to do both. SaaS developers can take data stored in Amazon S3, augment it with information stored in Amazon DynamoDB, then process with AWS machine learning services such as Amazon Textract or Amazon Comprehend to add new capabilities for their users. Security operations engineers can build reliable, observable, and auditable workflows that react to events from Amazon EventBridge then execute actions in Amazon EC2 to enforce IT controls. Mobile application developers can build a synchronous API that uses Amazon Personalize, Amazon Location Service, and Amazon Pinpoint to enrich the experience of their users.

We’re holding the Flex Your Skills Contest so you can share your best innovations for integrations across AWS Services. And with the recently released Workflow Studio in Step Functions, getting started has never been easier.

To enter, submit your working idea via a pull request to a public GitHub Repository here. Entries will be judged by a panel of AWS experts, and 3 winners will be selected. The Flex Your Skills Contest closes on 11/19/2021. Winners will be announced on 12/02/2021. This contest is open to residents of the United States only, but in the spirit of inclusion, we hope to expand to a global audience in the future. Submit your entry today!

Prizes

First Prize:

$500 in AWS Credits
A case study of your entry on the AWS Startups blog
An AWS Swag pack valued at $500
Announcement in the Serverless Office Hours session at AWS re:Invent 2021

Second Prize:

$200 in AWS Credits
An AWS Swag pack valued at $200
Announcement in the Serverless Office Hours session at AWS re:Invent

Third Prize:

$50 in AWS Credits
An AWS Swag pack valued at $50
Announcement in the Serverless Office Hours session at AWS re:Invent

Additional Prizes:

$25 in AWS Credits will be awarded to our 2 favorite entries per day.

Enter here, and see full contest terms and conditions.

Optimizing Logistics Channels with TheLorry and AWS

Mikey Tom — Tue, 19 Oct 2021 17:20:13 +0000

The rise of ecommerce has been well documented (just look at the skyrocketing stock of companies like Shopify). What’s sometimes lost in these conversations is the end-to-end coordination needed to support getting items from sellers to buyers. Amazon is well known for commanding its own fleet of vehicles to enable their marketplace around the world, but in some geographies, there is opportunity to define that customer experience—this is where TheLorry operates.

Founded in 2014, the startup has built a logistics platform that connects owners of lorries (similar to a cargo truck), vans, and trucks with customers in need of shipping. From consumers that are looking to clean house and get rid of unwanted goods to enterprises delivering large items like appliances, TheLorry is able to link them up with trucks to efficiently move what is needed.

“We have a few business models,” says Waqas Khalid Obeidy, Chief Innovation Officer. “Our B2C arm supports consumers who are moving houses, then we have our TheLorry for Business division that works directly with SMEs that sell large products, such as furniture or dishwashers, to ship them to purchasers. Lastly, we also operate a-less-than-truckload option, where we allow customers to ship just one item, and we piece together those orders into lorries that deliver them within 72 hours.”

A key component for each of the logistics startup’s businesses is route and load optimization—the company needs to know the most efficient way to move deliveries between points A, B, C, and so on down the road. Per Obeidy, this is where his team has had much success in leveraging AWS.

“When we started the company, the process was all manual. We’d take items to a warehouse and sort them by hand depending on where they’re headed. Drivers would then take what they can carry for each region. That would mean some of your packages might get delivered in one truck, while another comes later in another truck.”

To solve this, Obeidy and his team set out to build an AI-based SaaS service that could be applied to all three of their business units, improving route planning and, as a result, offering a better customer experience across the board. The results have been extremely promising, per Obeidy.

“We took a containerized approach in concert with AWS Lambda and Amazon ECS, making it as easily manageable and scalable. Each delivery has data points, like size, weight, or destination logged, which are then ingested by our system to optimize truck loading and route management. The results have been very promising, with the team seeing approximately 55% improvement in operational efficiency. For the drivers, this means less costs, less travel, more jobs. For the customer, this means we’re able to provide a faster and more reliable experience. For us, this means efficient operations, satisfied customers and a clean carbon-free environment”

This operational efficiency is much needed and timely, as the company saw a huge spike in business as a result of the COVID pandemic. TheLorry recorded roughly 15,000 items delivered in January of 2020—that number is now well over 130,000 each month and growing.

Obeidy and team’s optimization model was initially launched with the company’s operations in its hometown of Kuala Lumpur, but there are plans to further scale across all the cities the startup operates in across Southeast Asia. With massive growth and roughly $7.4 million in venture funding, TheLorry looks well positioned to continue bringing its customers what they need, when they need it.

How Stoke Therapeutics Is Turbo-Charging Drug Discovery Using Snakemake and AWS

AWS Editorial Team — Mon, 18 Oct 2021 19:56:22 +0000

By Eric Lim, Head of Bioinformatics at Stoke Therapeutics

As a biotechnology company, Stoke Therapeutics is dedicated to addressing the underlying cause of severe diseases by upregulating protein expression with RNA-based medicines. Identifying genomic signatures amendable to TANGO (Targeted Augmentation of Nuclear Gene Output), our proprietary approach to developing antisense oligonucleotides and selectively restore protein levels begins with computational analyses of both publicly-available archived data and privately-generated sequencing data. Through AWS, Stoke has quick access to computing resources without the active management of on-prem hardware, enabling us to close the gap between sequencing and interpretation and dedicate more time to science. Our proprietary bioinformatics pipelines have identified approximately 2,900 genetic diseases, which we believe to be amenable to TANGO.

Today, we’re sharing how we leverage AWS services, some scalability constraints we experienced in our journey, and the architecture that has worked well for us. To date, we have deployed this cloud-native infrastructure to analyze over 20,000 RNA sequencing samples with an estimated cost saving of over 70% when compared to a traditional approach using on-demand Amazon EC2 VMs with local EBS storage.

Overview of architectural design in AWS

CI/CD and workflow management in AWS CodeBuild

Our first program addressing Dravet syndrome, a severe and progressive genetic epilepsy, is in clinical settings. We are also pursuing a second haploinsufficient disease, autosomal dominant optic atrophy (ADOA), the most common inherited optic nerve disorder. Our bioinformatics workflows in analyzing genomics data are composed of several disjointed command-line utility tools that transform raw sequencing data into interpretable genomic signatures to identify the relevant disease and drug targets. Different tools often have their own set of inputs, outputs, resource footprints, and software requirements, so we use a workflow management system to manage versioning and dependencies between jobs, and most importantly, to ensure in silico reproducibility. At Stoke, we use Snakemake and conda/mamba to deploy analytical workflows within the AWS ecosystem of services. For more information, you can also review an incomplete but long list of existing workflow management systems used in bioinformatics.

Continuous Integration

Continuous Integration/Continuous Deployment (or Delivery) is the backbone of modern software and DevOps environments. Our goal is to transform hundreds of terabytes of genomics data into queryable scientific insights by automating as much as possible the deployment of our analytical workflows. The diagram below illustrates our approach.

We use CodeCommit as our Git repository. The team also subscribes to trunk-based development, where developers collaborate on a single branch, to avoid long-running branches from falling out of sync. Consistent with others, this methodology has substantially reduced many frustrating scenarios with merging code.

Once the code is committed and pushed into the master branch, AWS CodeBuild takes over to create a Docker image containing the pipeline definitions and install software requirements in the specified versions. Snakemake natively supports automated conda deployment via mamba (–conda-create-envs-only and –conda-frontend mamba) and a modest unit testing framework (–generate-unit-tests), increasing reproducibility. CodeBuild pushes the resulting image to the corresponding Amazon ECR registry that EC2 instances provisioned from AWS Batch will consume. Finally, we incorporate the quality control (QC) workflow as part of the CI pipeline to automatically assess data quality and store QC reports in S3.

Review

The majority of our sequencing data comes from the scientific literature. The reuse of publicly accessible data has immense potential for novel scientific discoveries, but there are often inherent quality differences between data sources. While deployment can be fully automated, we deemed that the data quality assessment is an important business decision. So, the team manually reviews QC reports for each genomic project and updates the corresponding metadata in Amazon DyanmoDB before triggering analytical deployments.

Continuous Delivery

The CI pipeline uses CodeBuild to automatically generate all necessary ECR images for the CD pipeline to deploy the analytical workflow to Batch. The workflow stores all intermediate data, essential log files, and results in S3. Results are converted into Parquet and compressed with Snappy. We utilize Amazon Athena and Amazon QuickSight as an interactive query service to analyze genomic results in a data lake using standard SQL.

Metadata in AWS DynamoDB

We use DynamoDB to store metadata of genomic samples. When DynamoDB receives a new item, the stream triggers a Lambda function to create a Batch job, which downloads the corresponding raw sequencing data from public repositories such as the Gene Expression Omnibus to S3.

Compute strategy in AWS Batch

Our compute approach is straightforward and predominantly uses AWS Batch to process genomics data. Batch is a fully managed service that enables us to scale our containerized genomic workflows to hundreds of thousands of compute nodes, all without actively managing the infrastructure. The diagram below illustrates our strategy.

A workflow may contain many independent container jobs. As each job may or may not run on the same instance, Snakemake offers several utility functions to facilitate the deployment.

We use S3 as a central storage service to exchange genomics data, results, and logs between containers. Snakemake natively supports remote I/O (–default-remote-provider S3 and –default-remote-prefix bucket_name). In addition, the functions also handle staging remote data before job execution, uploading outputs to S3 as job completes, and performing cleanup to free up disk space for jobs remaining in the queue.

We deploy the entire Snakemake workflow to Batch, which launches an on-demand EC2 instance serving as a head node to manage job dependencies and submission. Workloads defined in the Batch job queue autoscale the provision of Spot instances. The head node will run for the entire duration of the workflow, whereas each Spot instance will terminate when it is no longer processing jobs. We predominantly use 3-5 EC2 Spot instance types that provide an instance store, including R5d, M5d, and C5d. An instance store provides temporary block-level storage that is physically attached to the instance, and in our experience, produces superior I/O performance for genomics analyses. This approach eliminates using EBS as the compute storage, which further reduces costs.

Scalability

As a result, we have deployed this strategy to analyze over 20,000 RNA sequencing samples in a cost-effective and performant manner. However, as data grow exponentially and analyses inherently become more complex, we started hitting a few scalability constraints. Although entirely relying on instance store as the compute storage is both cost-saving and performant, the size of the instance store might not always offer enough capacity to accommodate every job. Without carefully planning for vCPU/GPU, memory, and storage requirements, jobs occasionally fail from running out of disk space.

Amazon FSx for Lustre provides transparent synchronization of data stored in S3 and is accessible as a shared filesystem across all instances. The service is attractive as most bioinformatics tools natively work with data stored on a POSIX-compliant filesystem. The team is currently evaluating FSx for Lustre as an alternative solution for genomic workflows that require large amounts of disk space per job.

Our strategy relies on S3 as a central storage location. Building job dependencies with many remote input and output I/O objects can be a time-consuming task. Some of our workflows take over several hours of processing before Batch deployment.

We implemented a custom cluster execution logic in Snakemake (–jobscript) to alleviate this constraint. When deploying a workflow, we instruct Snakemake to always run every job without first checking if the outputs already exist in S3. It allows us to delay interacting with remote I/O objects until the actual submission to Batch, a step we can parallelize and scale with ease. If the expected output of a job is already in S3, we interrupt Snakemake from submitting the job to Batch and record the job as complete. Instead of spending hours preprocessing all job dependencies of large workflows, this approach takes just minutes to start sending unprocessed jobs to the Batch queue.

Conclusion

Although Stoke is now a publicly traded and clinical-stage biotech company, our AWS journey began when we had just a few employees. AWS enables startups like us to stay agile and scale the compute infrastructure as we grow. Growth does come with pains, but we are continually improving and leveraging the latest technology in AWS to further our understanding of genetic diseases.

Eric Lim received his Ph.D. from Brown University in 2011. His work focused on using computational biology and high throughput genomics techniques to identify functional splicing elements in the genome. Eric has a decade of postdoctoral experience in the biotech industry and is currently the Head of Bioinformatics at Stoke. He leads a team of talented computational scientists in leveraging antisense oligonucleotides as a therapeutic modality to treat the underlying cause of genetic diseases. He has over 24 publications and patents in RNA splicing, and his work has led to numerous awards, including the 40 Under 40 in Cancer. Eric is also a member of the COVID-19 International Research Team.

Acknowledgements

The bioinformatics and IT teams have contributed substantially to the success in using AWS to process genomics data. I want to take this opportunity to acknowledge Sebastien Weyn, Jose Negron, and Stephen Yu.

Empowering the Blind – How .lumen Is Revolutionizing Blind Assistive Technologies Using AWS

AWS Editorial Team — Thu, 14 Oct 2021 13:45:36 +0000

Guide dogs have helped lead the blind for decades, and while the furry friends are an invaluable asset for many, they come with limitations. Training one can cost tens of thousands of dollars, and caring for them is a big responsibility. On top of that, they’re a scarce resource. There are 40 million blind people around the world—a number expected to jump to 100 million by 2050—but just an estimated 20,000 active guide dogs.

Canes can help close that gap to an extent. And other technological advances have made it increasingly possible for the visually impaired to consume and transfer information, like 3D sound maps and bluetooth beacons. “But when it comes specifically to mobility and orientation, getting out of the house and going to a new or unknown place, unfortunately, you only have the white cane and the guide dog,” says Cornel Amariei.

An extraordinary vision

Cornel, whose resume boasts a recent induction to the Global Business Hall of Fame and inclusion in Forbes’ 30 Under 30 Europe 2016 list, is the CEO and founder of .lumen, a Romanian startup working on adding a wearable device to that too-small list of mobility solutions for visually-impaired people. The company’s goal is to pack all the benefits of a guide dog into a headset, making getting around far easier for the millions of blind people who don’t have access to a trained canine.

It’s a goal that wasn’t possible as recently as 2013, when Cornel first had the idea to design a pair of glasses for the blind. Growing up around parents with locomotor disabilities and a sister with cerebral disabilities made him hyper-aware of both the possibilities and the limits of assistive technology, and he was eager to create a product that could make mobility more accessible.

Following his 2013 inspiration, Cornel played around with some product development for a mobility headset. His experimentation resulted in some promising lab results, including helping blind people catch a ball for the first time and pass through an open door without having first touched it. But it quickly became clear to Cornel that technology hadn’t advanced enough to make manufacturing a scalable device possible. That realization was dispiriting, but didn’t mark the end of Cornel’s hope for the future.

“We took it upon ourselves that the moment when the technology caught up, to actually build this thing. Then, in 2019, we realized the technology might actually be there, and we would be able to build what we knew we had to build,” he says.

Making meaning, not money

In January of 2020, Cornel began fundraising, and in May of 2020, he and a lean team of eight other employees had their first day at .lumen. That their inception came during a pandemic was a challenge, but not one they couldn’t overcome—and overcome quickly. Within a few months, they’d grown to 40 people and plan to be between 70 and 100 in another few months. No matter how big they grow, though, Cornel and his team are committed to one guiding principle.

“We founded .lumen with this idea of making meaning, not money,” Cornel says, but he and his team had tons of questions to answer before they could create a relatively small device that would pack the power and complexity of the human brain.

“There are a lot of things, which we take for granted, that are very simple for us as humans and incredibly complex to do in a system,” he explains. “We had quite a few unknowns. How well can we understand the world? What is relevant to the world to understand? The last and biggest one was, how will we represent that information to a blind individual quickly, reliably, and safely?”

An optical breakdown

They tackled that last question first, launching round after round of research to determine the best ways for the device to communicate information, beginning with existing methodologies guide dogs employed. Typically, when the animals are leading, they give their owners a gentle tug on the hand to steer them in a safe direction or alert them to barriers.

The .lumen device works in a similar way. Outfitted with five cameras that constantly take in the external surroundings the way a human or trained dog might, the device then computes that navigational information, contextualizes the nearby environment, and determines the information it needs to relay to the user—perhaps a nudge along a safe path, or a warning about an obstruction blocking the user’s way. Then, using a proprietary system of haptic and auditory feedback mechanisms, it conveys that information using a similar sensation to a guide dog’s tug on the hand, only to the user’s head.

With that development underway, the team started training models for production, but found the job too time-consuming with their previous provider—they had to wait months to complete testing for a single model.

“We want to have state-of-the-art machines, so we use neural network architectures, and these are very data-hungry,” says Caius Debucean, .lumen’s Machine Learning Lead. “If you want to train an AI model, it can take days, weeks, and even months.”

Tired of waiting, Caius insisted .lumen needed a more powerful cloud computing platform if it wanted to accelerate its research pace, produce a scalable model, and finally get an accessible mobility solution into the hands of millions of blind people.

Powered by AWS

AWS turned out to be the answer. Through NVIDIA Inception, a virtual accelerator that supports a global ecosystem of over 3,600 members, .lumen got access to AWS Activate, a startup-focused tiered program designed to accelerate long-term startup growth through the issuing of AWS credits, access to subject matter experts, and business and developer support. Activate gave .lumen the resources it needed to quickly get started on AWS and leverage its tools to grow.

“Now, we are training intensively on tens of different models. If you need to train those architectures and want to have state-of-the-art results, you need powerful cloud computation. This is where AWS has been big time for us, because they offered us one of the best products, one of the best services in terms of AI machines and AI services in order to train those models,” explains Caius. Not wanting to back up its data locally and risk corruption, the company also relies on AWS cloud storage to house the more than 10 million images and data sets used to train its AI system. “AWS provides very fast and very accessible storage programs,” he adds.

Thanks in part to that efficiency and accessible storage, .lumen’s development phase has gone even quicker than they anticipated. They’ve already been able to test their product with more than 200 blind individuals, even as pandemic protocol made that a greater challenge than it normally would be.

Going through that process and finally developing a product was an incredibly validating feeling for Cornel, who had been waiting years to see his idea turn into a viable product.

When asked about something they thought they needed but turns out they didn’t, Cornel’s answer is immediate: time. “Everyone told us it would take 10 years, and we did it in six months,” he says. “We can do everything – it’s all possible.”

Looking ahead

The team knows its challenges aren’t over. In addition to overseeing a production process, .lumen must figure out the best way to train people to use the device. They also anticipate needing to come up with solutions for connectivity or latency issues, so that as many people as possible can use their device, no matter where they are in the world. Business development could be another logistical hurdle, since the company wants to sell directly to governments or other providers that can distribute the device to blind individuals free of charge.

But the team isn’t worried that they’ll lose steam as they continue. For one, they’re confident that AWS will continue to provide the technical and business solutions they need to efficiently evolve and scale. Plus, they know they’ll never tire of their mission to design a product that has the potential to provide mobility on a scale previously unheard of.

“You look outside and you see that there are not so many solutions, and this is really what keeps me going,” says Caius. “Blind individuals are not empowered enough today in terms of mobility, and we are here to help millions.”

Travizory Leverages GitOps and AI to Help Countries Unlock Safe Travel in Just 4 Weeks

AWS Editorial Team — Wed, 13 Oct 2021 14:46:19 +0000

Guest post by: Barry McLaughlin – Senior Software Architect, Travizory and Vladyslav Shkola – Startup Solutions Architect

As the world begins to reconsider international travel, the challenge many governments now face is how to reopen borders to revive their economies, while keeping local communities safe and minimizing health risks. Lengthy lines and unfamiliar screening processes at many airports make it clear that existing systems simply can’t cope with the ‘new normal’ of traveling. Anticipating the need to navigate this incredibly complex and high-stakes landscape, border security experts Travizory developed a world-leading secure SaaS border security and management platform using cutting-edge biometrics, AI and machine learning technologies that enables countries to safely welcome visitors within a matter of weeks.

How the Seychelles opened their border in 2 weeks

Our Visitor Management Platform offers an easy to deploy, highly available, secure, system for governments around the world, including the idyllic islands of the Seychelles in the Indian Ocean.

In September 2020, Seychelles implemented a facial travel authorization system to streamline travel for visitors, requiring tourists to pre-submit information about their trip online, including hotel address, passport information, and vitally, proof of a negative PCR test before being granted permission to enter the Seychelles. Without this pre-authorization, airlines operating in Seychelles are instructed not to allow the passenger to board the flight.

The ability to risk assess and pre-approve passengers before arrival significantly reduces the burden on immigration officials and allows staff to focus only on the handful of cases that require additional attention. Easy access to rich, real-time data sets improves decision-making and planning as departments are able to quickly identify trends, e.g. increasing number of infected passengers, within the data.

Not only is this more efficient for staff behind-the-scenes, but it also helps to reduce delays from start to finish: at check-in, through to health and immigration checks on arrival. Crucially though, it also provides certainty to passengers that their holiday can go ahead without a hitch.

Finding a solution with AWS

The problem was clear: how could Travizory maintain the speed and agility of a startup, while working with regulated government customers? Classic software providers for governments could spend years in developing and months in deployment, but not Travizory. Despite a number of technical challenges that make this difficult to realize, by leveraging AWS’s vast ecosystem of services and on-demand infrastructure, Travizory is able to concentrate on what matters to their customers – providing safe and secure experiences to international travelers, and ensure fast deployment, high compliance and data protection standards, and delivering an always-on experience for travelers, especially during border checks.

Automating a deployment process with GitOps

At Travizory, one of the things that helps us to deploy fast is our GitOps approach. A term introduced by Weaveworks, it’s an operational model for Kubernetes (k8s) and other Cloud Native technologies which consists of a few main principles:

The entire system described declaratively
Git is a source of truth for the desired state of the whole system, not only application code, but infrastructure as well
Any approved changes to the state in git are automatically applied to a system
Infrastructure defined as a Code (e.g., Terraform, CDK or CloudFormation)

Travizory’s GitOps approach

Let’s dive deeper into how these principles are implemented in Travizory.

Declaratively: All our services are defined declaratively with Helm Charts and are deployed on Amazon EKS. Helm uses a packaging format called charts; a chart is a collection of files that describe a related set of Kubernetes resources. A single chart might be used to deploy something simple, like a memcached pod, or something complex, like a full web app stack with HTTP servers, databases, caches, and so on. With EKS, the Kubernetes control plane runs in an account managed by AWS, the Kubernetes API is exposed via the Amazon EKS endpoints associated with our cluster. Actual worker nodes run in our AWS account and connect to a control plane via the API server endpoint.
Git, source of truth: All the current code that defines our services and infrastructure is stored in Git. Developers push updates to individual components all the time. We have different Helm Charts stored in Git, per environment and per customer. This means that our infrastructure is always immediately reproducible based on the Git repository.
Any approved changes to the state in Git are automatically applied to a system: Once commit is pushed, a build process with AWS CodeBuild starts and the system is updated. That helps to maintain Operational Excellence, including monitoring and observability: we always know what is currently deployed. If a new system must deploy, it would be just a new branch for a new customer that will be automatically deployed. Currently we have a rolling update deployment with Helm and AWS CodeDeploy, firstly we deploy our code to the Test account and once all the health checks passed and QA team did the test, we deploy the same code into Production account. In the future we plan to have a Blue-Green Deployment, which is supported by the service.
Infrastructure defined as a Code: We use CloudFormation templates to automate the infrastructure deployment.

Why this approach

Our customers are governments for the most part, and we handle a lot of personal information regarding travelers. This means we need to be able to completely separate data for different customers – using different AWS accounts and separate Virtual Public Clouds (VPC).

To support this, we need to be able to reproduce everything in our stack: the infrastructure (k8s, elasticsearch, network config, etc), the security and monitoring, and our applications. Therefore, we employ a GitOps approach – where everything is code and everything is reproducible. It also allows us to know what is deployed where and to rollback when things go wrong.

Compliance

In highly-regulated industries, compliance is everything. This is where Travizory has benefitted from working with AWS on a so-called Shared Responsibility Model. AWS is responsible for the protection of the whole infrastructure that runs all the services and Travizory is responsible to keep their own services and data compliant and secure, leveraging managed services reduces the attack surface for our infrastructure and allows us to offload the undifferentiated heavy lifting.

Such features as encryption at rest and in transit, auto-expiration, and data retention policies of data help Travizory to keep highly sensitive customer data compliant and secure.

General Architecture

High-level architecture of Travizory

Our architecture consists of two main parts – main account and customer accounts. Main account consists of multiple CI/CD pipelines, general configurations, monitoring, development, testing and demo environments and Customer accounts are separate AWS accounts with Staging and Prod environments of our services

We are heavy users of Kubernetes, where all the services with business logic are located. In order to reduce costs, we evaluated Spot and Reserved instances and decided to go with Reserved due mainly to predictable demand per customer. By purchasing Reserved instances we were able to save up to 37% in comparison with standard on-demand instances.

Managed compute and database services

Amazon EKS helped Travizory to upgrade a major version of k8s on production with zero downtime, which would be much harder to achieve with self-managed k8s cluster.

The user data and application data is stored in Amazon Aurora, enabling speed and full text search features and scale up and down according to the demand and handle almost any load, be it just a few new customers or a new country.

Managed ML services

Heavy use of managed services, powered by AWS’s ML offerings, helps Travizory to develop features faster and deploy them in weeks or even days, instead of month or years, how it usually happens in environments with high compliance standards.

For instance, we are using managed ML services such as Amazon Rekognition, AWS’s image recognition AI to match user photos to passport photos, to get likelihood percentage of facial match as well as suggest whether a selfie is a good match as passport photo; Amazon Textract allows us to process customer documents and extract information, such as date and test result from the COVID certificates, Amazon Comprehend and its NLP services allow the system to extracts additional information from documents: entities, such as places, people, locations and personally identifiable information (PII), which are used to fill up a tourist profile and maintain our data quality

Conclusion

In just 4 weeks, Travizory enabled local authorities to safely and securely reopen to international tourists. Since March 2021, numbers of tourists have shot back up to 50% of pre-COVID levels. Our flexible and agile approach allows us to reopen borders almost immediately, so that travelers can seamlessly visit destinations and tourist dependent countries can support their healthy economies. Working with AWS, Travizory is paving the way for what post-COVID travel can and should look like, safe, seamless and fast.

	Barry McLaughlin is Senior Software Architect with Travizory. He aims to be always learning and trying new thing, with a dedication to continuously improving our software. Currently focused on cloud based technologies, Kubernetes, CI/CD pipelines and anything that helps make the boat go faster.
	Vladyslav Shkola is a Startup Solution Architect at Amazon Web Services (AWS) where he is helping startups to make sure that they can deploy the best, most scalable, and most secure architectures possible. Vladyslav has a Master degree in Distributed Systems Engineering.

Building Community with Common Room and AWS

Mikey Tom — Mon, 11 Oct 2021 19:18:40 +0000

Tom Kleinpeter, co-founder and Chief Architect

Communities are vital to the health of individuals and companies, but they can be sprawling, disjointed, and difficult to grasp—especially online. Enter Common Room, which co-founder and Chief Architect Tom Kleinpeter describes as a community-intelligence platform that provides a single view into everything that’s important in your online community, across all the different places it might be happening.

Community exchanges—from complaints, to requests for help, to feedback—often occur across many separate platforms, like Twitter, Slack, Intercom, Shopify, and more. Common Room compiles these interactions and mines them for what matters most, providing a deeper and clearer perspective on the health and metrics of a community, and the individuals or organizations that are engaging with it.

“If somebody has 20 minutes a week to interact with their community, how can you make sure that they’re seeing the most important feedback from their customers?” That’s the open question Kleinpeter and team are looking to answer with Common Room.

Many of the startup’s customers are developer-focused companies, like Imply, a provider of data analytics solutions. “Historically, developer tools are just complicated, so having a community of people that can talk about it and help each other is particularly valuable,” explains Kleinpeter. But they also work with companies outside the developer space—like Figma, a design tool with a rich online community.

Common Room sorts through an immense and diverse array of data—Tweets, GitHub comments, Slack messages—to homogenize and unify everything into one place. This presents unique engineering challenges, at the heart of which is the issue of “surfacing signal and meaning from the community,” as Kleinpeter puts it. AWS machine learning services facilitate this process.

Example view into Common Room’s dashboard

Amazon Comprehend’s sentiment analysis allows Common Room to identify not only the negative comments in the community—who’s angry? who’s upset?—but also the positive feedback. For example, “If you wanted to make a wall of love of all the best stuff you’ve gotten from your community, you can and should do that,” says Kleinpeter, “That’s super motivating for people at a company.” As Common Room looks to the future, they’ve started experimenting with custom classification and key phrase extraction.

Common Room also relies on other AWS services like Amazon RDS, Amazon S3, and SQS, technology that Kleinpeter appreciates having “just an API call away.” And using AWS Fargate to run containerized applications allows Kleinpeter to keep the company’s developer and production environments exactly the same, while making scaling things up very easy. Security is also much simpler than it would be on a physical machine. These are key features for Kleinpeter, who views “building things as simply as we can” to be a central facet of his job. “Early on, we wanted to pick a simple architecture that both provided guardrails and was simple enough for everyone to understand.”

This streamlined architecture provides the foundation for Common Room to ultimately create easier-to-understand views of communities for its customers. “This is the first step—visibility. I think there’s probably a lot more we can do going forward,” says Kleinpeter. He envisions Common Room building on their current work, finding new ways to foster connections, and becoming an indispensable tool to run communities.

And the company looks to be well positioned to do just that. With roughly $52 million of venture funding in the bank, which includes a sizable $32.3 million Series B in April 2021, team Common Room has plenty of fuel to propel continued growth.

BlackBuck builds future-ready Digital Freight Marketplace using AWS

AWS Editorial Team — Thu, 07 Oct 2021 20:27:17 +0000

by Rahul Aggarwal, Director of Engineering at BlackBuck, and Venkatramana Ameth Achar, Solutions Architect at AWS

BlackBuck, India’s largest trucking platform, is a digital freight marketplace for shippers and truckers to conveniently discover each other, providing services such as FASTag (an electronic toll collection system), fuel cards, GPS devices, and insurance, among others, to efficiently manage their fleet. BlackBuck’s business and users have grown rapidly from a few thousand users on the platform to more than 1,000,000 users. With the goal of becoming the world’s largest technology-driven trucking platform, maintaining a data-driven approach, as well as strategic product improvements, put Blackbuck well on its way.

Use Cases

BlackBuck started as a freight marketplace and later expanded into managed services, a shift that brought new sources and formats of data. Data stakeholders include:

Sales: Our sales team identifies potential demand and supply clusters from various data patterns in clickstreams, order booking search, location data spatial clusters, goods types, and truck types.
Operations: Geofence alerts provide data for Turn Around Time (TAT) analysis of different stages in operations workflows, like order fulfillment to improve user experience.
Finance and Pricing: Decisions on dynamic pricing and bid pricing for the freight marketplace are based on historical pricing data, analysis of user activities, liquidity based on real-time clickstream data, and transactions data.

Revolutionizing Architecture with Microservices

BlackBuck’s technology stack has evolved from a monolithic architecture to 100+ microservices and platform components to meet performance, scalability, cost, and feature needs. Each microservice can have different technology choices and nuances around data storage and retrieval, hence it became increasingly complex to build data-centric applications. BlackBuck needed to build a platform that would enable it to quickly build and deploy data heavy and machine learning (ML)-based applications.

The former architecture used a self-managed Presto cluster with read replicas of databases from different applications. However, this structure was difficult to scale, requiring costly additional Presto nodes or upsizing read replicas. It also lacked support for streaming use cases, third-party data ingestion capabilities, or governance around schema upgrades. Changes would also need to be made to support cost attribution and data science use cases, such as feature generation, model training, and orchestration.

New Data Platform architecture

BlackBuck set out to build the new platform with clear design considerations in mind: supporting ease of data ingestion and with strict governance, an open data lake with fact generation and high query performance, real-time stream processing, and avoiding any cost surprises.

Data ingestion with ease and with strict governance:
- Data Platform API Server and Schema Registry: Incoming data schemas are registered in self-managed Schema Registry Service before ingesting data into Kafka. Schema Registry also maintains schema versions to ensure changes are backward compatible. Validity of each data record is checked using the schema id present in the record, invalid data record is rejected to maintain data sanity. Rate limiting per schema was put in place to ensure optimal usage of storage capacity, especially when users with large data sets were onboarded onto the new platform.
- Data Platform SDKs: SDKs were built for backend microservices and mobile apps to send transactional logs and custom events into the platform. The SDKs handled batching, and provided data delivery guarantees against infrastructure failures.
- Change Data Capture (CDC) using Debezium: Kafka cluster with Debezium connectors are used to get new data from Amazon RDS and MongoDB into Kafka. In case of schema changes in Kafka, a AWS Glue Crawler would be triggered to update AWS Glue Data Catalog.
Open data lake, fact generation, and high query performance: Parquet was selected as the best fit for its data storage format. For the query engine, Amazon Athena was a winner from a management and cost perspective. AWS Glue jobs merge small files in a single partition to further improve query performance. Creating a Table from Query Results (CTAS) queries are scheduled to generate reusable facts. CTAS creates a new table populated with the results of a SELECT query.
Real-time stream processing: With the data streaming through Kafka in real time, real-time stream processing use cases were easy to solve. For example, we used windowing functions in Apache Flink on Amazon EMR to solve a use case where we need to send custom notifications to users based search patterns. Data at rest in the data lake is also used in the Flink jobs when needed.
Cost controls and data security: Amazon Athena with Parquet data on Amazon S3 provided up to 10x performance improvement over our former setup. However, cost became a concern, as users stopped optimizing queries because they were getting results quickly. Also, at times, users needed to query a huge data set on a non-partitioned key. Different Athena workgroups were created for different teams to attribute query costs to respective use cases. Athena workgroups was configured to cancel queries breaching thresholds, alert on daily limit breaches, and restrict user access to specific data.

The next step was to build a data science platform that enabled the data science team to explore the data from different sources, and quickly build, train, and deploy models into production.

Data Science Platform Architecture

Feature generation, feature store and model training: Flink on Amazon EMR is used for real-time feature generation. The Data Science team uses self-managed Apache Airflow on Amazon EC2 to schedule queries on Amazon EMR. The features generated are stored in Cassandra, and are cached in Amazon ElastiCache for Redis with a TTL for reuse.
Model Build: We needed a model development environment which supports a variety of ML libraries to assist with deploying models and provide freedom to build models using language and libraries of the team’s choice. After evaluating their options, the team landed on MLflow on EC2 for its end-to-end lifecycle management.
Model Deployment, Wrapper Service, and Experimentation: Amazon S3 was used as a model repository. When a model is promoted to production for inference, a CI/CD pipeline creates a Docker image and registers it into Amazon Elastic Container Registry, and then deploys it as an API endpoint on Amazon SageMaker. Amazon SageMaker supports high performance machine learning inference. A wrapper service was used to abstract client applications from feature sets and models used. The wrapper service internally takes cares of looking up the features from the feature store and use the relevant model on SageMaker. This also enabled A/B experiments on different models for the same business use case.
Monitoring: All the model inferences and key performance metrics are pushed to the Data Platform for further analysis and alerting. The results of A/B experiments are analyzed by joining these key metrics with actual transactions on the platform and understanding each model’s performance.

BlackBuck’s Supply App

With the Data Science Platform established, fleet owner could now search for loads using Supply App.


Search Dialog Interface	Search Results Interface

The major source of data to figure out relevancy for a user are the click streams from the users’ activity on the app, location data from FASTag swipes and GPS devices, bookmarked favorite routes by the users, and the transactions done by the user on the platform. Learning to Rank (LTR) supervised ML model is used to solve ranking problems. For LTR, we use a Discounted Cumulative Gain (DCG) based cost function, which is a measure of ranking quality, measures the usefulness, or gain of a load on its position in the result list. The result is further refined based on user’s historical and real-time activity. When a user searches for loads or opens the find loads sections, an API call is made to Amazon SageMaker. The inference from SageMaker is used to show ranked loads on the app.

Conclusion

The new scalable data platform and data science platform gives BlackBuck a unified view of all our data and makes it easier for us to build and deploy analytical, reporting, and ML use cases. As a result of this implementation:

We achieved over 90% compression over raw data using Parquet columnar storage Data lake on Amazon S3, reducing storage costs.
We saw 10x performance improvement query performance compared to existing setup in most of our queries.
We reduced overall costs by 50% compared to when all queries ran on raw data.
Quick experiments with multiple models on SageMaker helped the Data Science team with A/B testing of models without the need for application changes.

Looking ahead, real-time querying, building and training models at scale, and rich visualizations and dashboards will continue to help BlackBuck use open-source technologies to take data and delight our customers.

Rahul enjoys working closely with Business and Product to solve their key problems using technology. He spends most of his time designing new systems and likes to evaluate and adopt new technologies while paying close attention to low-level details, scalability, performance, operational efficiency and cost. Outside of work, he spends his time playing online games and listening to different genres of music.

Ameth is a Solutions Architect at AWS. In his 20+ years of work experience he has helped customers achieve their business objectives using his skills in systems integration, prototyping, process automation, architecting solutions on cloud and data centers, as well as team management. He has served customers in ecommerce, semiconductor, finance, education, telecom, aggregators, logistics, media, gaming, security, and electric vehicles domains. On a personal front, he continues to improve his knowledge on renewable energy, music, and digital photography.

Catching Up with the Winners of the 2020 AWS University Startup Competition

Mikey Tom — Wed, 29 Sep 2021 17:56:32 +0000

Universities have long been at the center of startup lore. Red Bull, pizza, late nights coding, it’s all synonymous with the famous founding stories for many of today’s tech giants—I mean, we’ve all seen The Social Network.

And it makes sense. College students often take novel approaches to existing problems, may have a higher risk tolerance than those later in their careers, and are surrounded by other motivated people looking to collaborate. With all that in mind, AWS launched its annual U.S. University Startup Competition in 2020 to find and support the best entrepreneurs currently building while also juggling their studies.

Last year, over 600 startups applied from 185 different U.S. universities. Applications were reviewed by representatives from the AWS Startup Business Development team, who then selected 10 teams to compete in the last round. Each startup was paired with a subject matter expert from AWS to help them polish their pitches before their final presentations. Winners received up to $20,000 in cash and up to $100,000 in AWS credits, as well as intros to AWS partners like Techstars and Dorm Room Fund.

In preparation for this year’s edition of the competition, which is currently accepting applications, we caught up with last year’s finalists to get an update on their progress. Check out the winners from last year, and apply here if you’re interested in participating this year.

Big & Mini

Anthony Zhou, Allen Zhou, and Aditi Merchant

Having taken home first place in last year’s competition, the team at Big & Mini has since been heads down working to scale their platform that connects college students and senior citizens for meaningful conversation and mentorship. The idea was birthed out of the pandemic, when many people were feeling isolated. What started at the University of Texas at Austin as a simple spreadsheet has now evolved into a robust service, and the team even recently secured a partnership with AARP to bring Big & Mini to many of their chapters across the US.

“The AWS University Startup Competition was great in that it exposed us to a handful of mentors that have been really helpful as we’ve progressed. We’re still in touch with those connections and lean on them for advice today.” – Aditi Merchant, co-founder

Stratodyne

Ed Ge, founder & CEO

The now team of five behind Stratodyne are building blimps for surveillance in conflict zones around the world. Per founder and CEO Ed Ge, they had originally targeted the agriculture industry, but found that farmers weren’t open to spending the amount of money that would make for a viable business. Stratodyne then moved their attention to the security and defense space, where they were able to secure a contract with an organization in Nigeria to monitor the terrorist group Boko Haram. Ed founded the startup while in school at the University of Missouri-Columbia, but has taken a leave to focus on building.

“This next year will be focused on delivering on our contract in Nigeria, with hopes that we can use that as a case study to acquire more customers down the line. The AWS University Competition was really helpful in our early days—the capital from placing in the finals went directly to kickstarting things and the internal AWS connections assisted in building the software behind our AI-powered blimps.” – Ed G, founder & CEO

SwiftSku

Mit Patel and Daniel Mazur

There are roughly 153,000 convenience stores in the United States, collectively representing a $650 billion industry, according to SwiftSku co-founder and CEO Mit Patel. The crazy part? Much of the bookkeeping at these stores are done manually with pen and paper, or with antiquated records systems. Not only does that take a lot of time, it’s ripe for error. Patel felt that pressure first hand growing up while helping his parents manage their own convenience store. Out of that experience was birthed SwiftSku, a startup that Patel launched with his co-founder Daniel Mazur while they attended Auburn University. The premise is fairly simple: SwiftSku plugs into existing point-of-sales systems to help store operators optimize their inventory via analytics, automated reporting, and dashboards, offering an easy look into the current status of everything. The team has since raised roughly $3.2 million in funding and recently graduated from the prestigious Y Combinator accelerator program.

“The AWS University Competition was a great way for us to meet fellow founders who were going through similar struggles trying to build an early-stage startup. Since then, we’ve grown from roughly 60 stores to being on track to hit our goal of 5,000 by May 2022.” – Mit Patel, co-founder & CEO

Skoop

Josh Cooper

We’ve all seen physical billboards and advertising displays, but a growing segment of the advertising market has been moving to digital versions, for obvious reasons. They’re more dynamic, eye-catching, and can rotate between multiple messages. Skoop, a startup founded by Josh Cooper while at Michigan State University, is looking to capitalize on that opportunity by building a platform for digital advertisers to easily manage these billboards. Per Cooper, there are currently 25 million billboards in the United States, and it’s thought that 50% will be digital in the coming years.

“One of the main benefits we saw from the competition was the prize of $100,000 in credits, which enabled us to test a variety of services and build on AWS.” – Josh Cooper, founder & CEO

Are you ready to join their ranks? Submit your entry to the AWS US University competition by November 1 for your chance at up to $20K in cash and up to $100K in AWS credits!

Shipper’s Digital Transformation Journey to Build Next Generation 4PL Supply Chain

AWS Editorial Team — Mon, 20 Sep 2021 23:11:59 +0000

Guest post by Marvinus Kokoh Arif, CTO, Shipper, and Riza Saputra, Solutions Architect, AWS

Shipper is one of the fastest growing technology companies in Indonesia, providing end-to-end digital supply chain solutions for businesses of all sizes in a market where last mile fulfillment and delivery is estimated at US$80B. Via our proprietary software, we connect the country’s largest tech-enabled logistics network of third-party logistics providers (3PL), first-mile delivery agents, micro-fulfillment hubs (e.g. mom-and-pop shops), and over 222 warehouses across 35 cities in the country.

In this blog post, we would like to share our technology transformation journey and how we have grown and supported a couple hundred orders per day back in 2017 up to hundreds of thousands orders per day in recent years with the help of Amazon Web Services (AWS) managed services such as Amazon Elastic Kubernetes Service (Amazon EKS).

In early years, our architecture was as simple as providing enough compute power for our applications. The focus was really to prove out our business model and significant cost controls from an infrastructure perspective.

Early architecture

We began to realize that as our business grew, so did our problems, including extended server outage, slow application loading, and extensive manual infrastructure support from the team. Insufficient backup strategy also caused our systems to be vulnerable and resulted in long recovery issues when problems arose.

Environment Separation

In order to alleviate some of the problems above, we started to take advantage of AWS managed services such as Amazon Relational Database Service (Amazon RDS), which helped us tremendously for patching, scaling, as well as performance tuning by utilizing read replicas. Servers were segregated by capabilities, nightly backup, and environment separation to support more rapid development.

Environment separation

At first, things were looking a bit better, but soon enough, we realized there were several fundamental issues which we were still facing such as inability to scale out, untraceable issues, lack of logs monitoring and metrics, and the fact that all the applications were still monolithic.

Breaking into microservices

As our business continues to grow rapidly, we felt that rearchitecting was needed for long term investment, from a scalability, reliability, and cost optimization perspective. We identified some key areas to modify:

Isolate service per domain
Build microservices not nano services
Keep API and event driven architecture in mind
Invest in automation
Implement stress test and performance monitoring tools
Avoid single point of failure and scale horizontally

Microservices architecture

We began by breaking our monolithic service into several microservices based on their responsibilities. A common pattern is the strangler pattern, where we incrementally separated specific pieces of functionality from the legacy application into new microservices. During this process, our legacy applications were still running and fully functional while all new components were containerized and deployed with Amazon EKS.

Amazon EKS makes it easy to run Kubernetes and simplifies our architecture by reducing complexity (~60%) and cost saving (~30%). By using Amazon EKS, our applications were able to scale up horizontally and support self-healing deployment. Within the Kubernetes nodes, we ran many microservices segregated by their domain services. Going forward, containers become de facto of our development paradigm without worrying about complexity of the orchestration layer.

We also implemented a data replication via CDC (Change Data Capture) process where we synced data from existing repositories into new repositories: Amazon RDS and Amazon Opensearch Service. One of the biggest benefits we saw was a huge search performance improvement via Amazon Opensearch Service versus querying the database directly. We were able to offload traffic from our main database and at the same time, we were replicating data from legacy to new without system interruption.

For our warehouse solutions (not depicted in the diagram above), we chose Amazon Aurora as our database engine of choice since our warehouses increased exponentially. With Amazon Aurora, replication across multiple availability zones keeps our application transparently recovered when failures occur.

Conclusion

We’ve learned many lessons from this journey but feel like we’re just getting started. We’ve had to change our mindset and tailor the solution based on our business needs. By using AWS managed services, we saw lots of improvements especially from a scalability, availability, and cost optimization perspective. The more we grew, the more complex our architecture became, but it was well worth the investment. We hope our story with Amazon EKS, Amazon Opensearch Service, and Amazon Aurora can help inspire you in your own journey as well.

Build your startup’s infrastructure in minutes with Build on AWS

AWS Editorial Team — Wed, 15 Sep 2021 18:30:03 +0000

We’re excited to announce the launch of Build on AWS, a new offering from AWS Activate designed to help startups build their infrastructure on AWS in minutes. Build on AWS is a collection of infrastructure templates and reference architectures covering a wide variety of solutions curated specifically for startups. These solutions are built by experts at AWS and based on AWS best practices.

Startups are eager to get started on AWS, but often ask, what’s the best infrastructure for my use case? And while experimentation is often central to creation, startups want to start building the foundations of their MVP as quickly and efficiently as possible. With the launch of Build on AWS, we’ve simplified the first steps of launching scalable, reliable, secure, and optimized infrastructure tailored to startups’ industry or use case. This enables startups to focus on building their core product knowing they’re using AWS best practices for their underlying cloud infrastructure. With hundreds of deployments, every template is ready for production traffic. Startups can also rate the solution templates to help other startups, and request new templates—all from the Activate Console. Startups who need additional support in customizing configurations can also reach out to AWS Support or AWS IQ.

Build on AWS features CloudFormation templates and reference architectures. CloudFormation templates are deployable with a single click, and reference architectures contain an architecture diagram that can be replicated. Within Build on AWS, startups will find basic solutions and personalized solutions. Basic solutions contain core solutions to get started on AWS, including Hosting a WordPress Website on Amazon Lightsail and Building a Data Processing API using Serverless. A guided path is also provided to help startups choose and deploy the right solution. Personalized solutions are solution templates that are relevant to the startups’ industry, interests, and AWS usage. Startups also have the ability to search through all of the templates by use-case or by underlying AWS services.

Start building today

Log in to the AWS Activate Console and navigate to the Build on AWS tab. You can also watch a quick introduction to Build on AWS that includes a brief demo on how to access the templates. Reach out and let us know what you think via the Activate Console, especially if you have recommendations for new templates.

If you’re not a member of AWS Activate, Activate is our startup program that has helped thousands of startups become the household names we all know and love including Airbnb and Slack. Apply for Activate to receive free AWS credits, technical expertise, architecture guidance, training, help with go to market and so much more.

We can’t wait to see what startups build on AWS.

Carsome: Leveraging Automatic Car Plate Masking on Amazon SageMaker to Focus on Growth

AWS Editorial Team — Fri, 10 Sep 2021 00:07:02 +0000

Guest post by Ammar Syatbi, Data Scientist, Fares Hasan, Data Scientist, Piyush Palkar, Chief Data Officer, Carsome

Carsome is Southeast Asia’s largest integrated car ecommerce platform. With operations across Malaysia, Indonesia, Thailand, and Singapore, we aim to digitize the region’s used car industry by reshaping and elevating the car buying and selling experience. We provide end-to-end solutions to consumers and used car dealers, from car inspection to ownership transfer to financing, promising a service that is trusted, convenient and efficient. Carsome currently transacts around 100,000 cars annually and has more than 2,000 employees across all our offices.

The value Carsome wants to achieve by innovation in this sector goes in two ways. First, it’s important for us that our customers have an excellent experience when they use our service, so the innovation we bring must keep the good work up and carry on making it better. Second, our operations will be more efficient and evolve to manage our growth plans.

Streamlining inspections with machine learning

At Carsome, we understand the importance of thorough checks to offer the best prices and build buyer confidence. Typically, a car takes up to 30 minutes of inspection where our professional inspector will generate 175 points and take pictures annotating various notes about the car condition and appearance. As our business has grown rapidly, it’s critical to streamline this process while maintaining high quality.

Inspection is a strategic step in the car journey, and we are aware of the sheer innovation this step deserves to pioneer and transform the industry. For instance, among the images taken will be front and back images of the car where the inspector has to manually blur the car plate numbers to preserve the privacy of our customers. This small part takes time though a few minutes only but it adds up when you have 100 cars and execution takes place on the small screen of a phone, which is less than ideal.

There are several ongoing efforts to transform this segment of the journey leveraging artificial intelligence (AI) and machine learning (ML). The team has proposed automation of the process with deep learning, completely freeing the inspector from the arduous work of blurring the car plate numbers from images manually and saving the time spent doing this. This will increase the inspection capacity and efficiency in terms of the number of cars the inspector can inspect. There have been several use cases on this front and our small prototype propelled the plans to go ahead and bring artificial intelligence into the center of the inspection process.

However, due to Carsome growing rapidly, it was challenging to predict the adequate compute resources needed for this project deployment. We needed something that is easy to scale up and scale down depending on market demand and operation plans. The conventional method of building everything directly into a vanilla Amazon EC2 instance cannot adequately address the needs. Though we can manage the resources efficiently using Amazon EKS, we needed a solution that is quick and easy to set up and maintain so we can focus more on data science and machine learning instead of managing the infrastructure.

Solution

We ran our experiment on Amazon SageMaker, which is a fully managed service for the whole workflow from training a deep neural network model into inference. It provides us with just what we need. SageMaker helps data scientists and developers to prepare, build, train, and deploy ML models quickly by bringing together a broad set of capabilities purpose-built for ML. We can build, train, tune, and deploy our model without having to think much about managing the infrastructure. The diagram below shows the high-level architecture.

Figure 1: High level model training architecture

The evaluation of the results can be carried out in multiple ways based on the use case. For this particular case, we have used the SageMaker Batch Transform to run an evaluation for the model performance. The evaluation that’s carried out looked at both aspects of the accuracy and the actual image masking quality. Since the model is stored in S3, we can always come back to it when we need to or in the case of comparing improvements.

For our modeling approach, we initially tried YOLOv3 to quickly streamline the process and build a successful test run using Tensorflow framework. YOLOv3 is a real-time object detection system. It takes the matrix representation of the image as input and generates a list of detected objects in the image, along with their coordinates and bounding boxes. In our context, the image refers to car images and the objects refer to the car plates. The image below illustrates some of the use cases of the YOLOv3 being limited in producing aesthetically pleasing results, which was the reason for us exploring another algorithm.

Figure 2: Result samples for YOLOv3 model

You can look at the plate area and recognize the following limitations:

● Inability to scale to car plate number size without masking neighboring regions that are not part of the plate.

● Inability to work with diagonally aligned objects which is a problem because that’s how most people take pictures naturally.

The limitations in the YOLOv3 model were obvious based on the aesthetic of the results we have obtained, which were always in bounding boxes instead of polygonal with unconstrained orientation. This means that we have to use a model that provides better results. We decided to adopt the Mask R-CNN (Masked Region Based Convolutional Neural Network), as it is proven to excel in solving such problems. In our context, Mask R-CNN goes a step above the bounding boxes into segmentation based output which has proven to be solving the limitations we observed working with YOLOv3.

Since Mask R-CNN is not among the built-in algorithms in SageMaker yet, there comes the need to build a custom container for it. In our case, we prefer using our own container image since we find it easier to adopt the existing code with many dependencies rather than using SageMaker prebuilt container images. The custom container feature adds value to us where some of the models might be new or innovated by teams working on various problems. Having the ability to build customized containers for them to be orchestrated in SageMaker is a real advantage. The following image compares the result from both models for the same car and you will see how Mask R-CNN scores high in all aspects of comparison. You almost will think the images were edited by software.

Figure 3: Car plate blurring comparison between YOLOv3 and Mask R-CNN

The image on the left is masked by YOLOv3 while the image on right is masked by Mask R-CNN. You can see the difference in the size of the blurred region and that precision and the ability to zoom into pieces and details are achieved by using Mask R-CNN segmentation capabilities.

SageMaker provides us with various options for deploying the model for inference at scale including SageMaker Hosting Services, SageMaker Asynchronous Inference, and SageMaker Batch Transform. The SageMaker Hosting Services simply means that we will have a persistent endpoint that will receive masking requests at any time and fulfill those requests. For our use case, we opted for SageMaker Batch Transform. We don’t need the car plate to be processed in real-time. Furthermore, we can save costs by leveraging the pay-as-you-go pricing of the SageMaker Batch Transform, which is based on the duration of the resources being used. Batch processing moves along with our operations so we can respond in time for requests and support business in sufficient time.

Workflow & performance

Leveraging our automation infrastructure and Apache Airflow orchestration we have built the workflow as a Directed Acyclic Graph (or DAG). The task starts with reading all car images that will be due in the next window and preparing them for the model. Since the workflow is scheduled, we trigger it at times agreed upon beforehand with our business and product stakeholders. So batch transformation will be invoked and all the images qualified will be masked and stored to be available for retrieval at any time needed. The diagram below shows the inference pipeline architecture.

Figure 4: The inference pipeline architecture

We have tested the model on Malaysian car plates. Over 100 car images, 99 of the images were correctly blurred. Only one image showed a slight issue as a result of the sun glare in the background which made the car plate region dim. However, such a case is niche and in our research, we only found one image taken in the aforementioned condition. Overall performance is impressive and we are confident in the results obtained.

The impact of such a solution in terms of time is spectacular. We have allowed inspectors to perform more by trimming time off such tasks. In our moderate estimations, our inspector spends at least 1 minute masking the plate number of a car (times evaluated in the ground based on 2 images per car performed by an experienced inspector). On the other hand, our automated system takes 2.4 seconds to mask images of one car. To put this in perspective we have reduced processing time by approximately 29x.

Future

Growth in Carsome is multiplying and therefore, we are building systems that multiply human productivity. We have a wide range of problems to solve via deep learning and artificial intelligence to improve our products and services. Our dealer auction recommendation engine went live earlier this year and with it, we set the tone of innovation with data. The vision is to embed intelligence in every stage of the business operations. This is our first round of optimizing the inspection process and we aim to build diagnostics that aid the human inspector.

Carsome’s operations focus on 3 countries: Malaysia, Indonesia, and Thailand. This makes it a requirement for any solution we build to be able to perform at scale in all 3 countries. Operating in these regions with the potential to grow beyond brings about the locality challenge where the same solution needs to be tweaked further before it can work on a different country or region. It’s always the case where we build solutions locally with the intention of deploying this solution regionally to touch every business branch and reap the positive impacts in all our operations.

The time is ticking for us to bring this solution to our regional operations and that is our priority now. This solution gave us a long view of the opportunities we can bring into our process and how AWS services like SageMaker aids in reducing the complexity often faced when building machine/deep learning solutions.

Author Biographies

Ammar Syatbi is a Data Scientist at Carsome. He started as a Software Engineer to be a full-stack Data Scientist leveraging deep neural networks in providing a business solution. He always has Python for breakfast and loves solving problems with machine learning.

Fares Hasan is an impact-driven lead data scientist with an affinity for building recommendation engines and customer-facing data products. Enjoys working with startups and building pioneering data teams.

Piyush Palkar is a data thought leader and has an experience of covering a range of areas and responsibilities, both technical and non-technical in nature in order to create and execute data & analytics programs to drive business value and embed data driven innovations in products. He holds the position of Chief Data Officer at Carsome and heads 4 departments viz. Business Intelligence, Data Science & Advanced Analytics, Insights, and Data Engineering. He is also the Country CDO Ambassador for the nation of Malaysia as part of the global partnership representing the Massachusetts Institute of Technology CDOIQ, the International Society of Chief Data Officers, Institute for Chief Data Officers, and CDO Magazine.

SaaS Founder Series: Clumio CEO on Steering the Startup that Raised $11M Pre-Product

AWS Editorial Team — Mon, 30 Aug 2021 23:06:09 +0000

Guest post by Afza Wajid, Mark Birch, and Zia Rehman, AWS

The growing customer preference for software-as-a-service (SaaS) puts enterprise software startups in a rare position of advantage compared to established firms. AWS SaaS Factory invited Poojan Kumar, CEO and Co-founder of Clumio, to share the early successes and learnings from steering the startup that is disrupting a segment filled with corporate behemoths.

Having clarity of vision and staying true to it

Poojan Kumar, CEO & Co-founder, Clumio

“Gone are the days of shipping a piece of software,” says Kumar, emphasizing that today, software must be delivered “as a service.” Customers’ widespread preference of transitioning to cloud-based SaaS offerings has resulted in a “big discontinuity” between the positioning of many market leaders in enterprise software and their customers, creating a promising opportunity for new players. The scale of this opportunity inspired Kumar and his co-founders to launch Clumio in 2017.

A serial entrepreneur with over 20 years experience in enterprise software, Kumar has a passion for tackling complex computer science problems relevant to large growing markets, that, when successfully solved, result in valuable intellectual property (IP) and drive business growth. But such problems take significant time and effort, and Clumio’s founding team recognized early that even getting to a minimum viable product (MVP) would take significant investment. “We didn’t write a single line of code until we raised the Series A funding,” the CEO reflects, because an unsuccessful funding bid would be viewed as an early indicator that the business proposal needed refinement. In October 2017, Clumio successfully raised $11M in a Series A led by founding investor Mike Speiser, Managing Director at Sutter Hill Ventures. Today, the fast-growing startup has raised a total of $186M in funding and has grown to over 100 employees in less than four years since its founding.

After raising its Series A round, Clumio launched its first SaaS offering, Clumio Protect, a fully-managed, secure-backup-as-a-service platform, hosted on AWS. The award-winning Clumio Protect is a cloud-native, multi-tenant service that delivers comprehensive data protection for enterprise workloads. Coupled with reporting for compliance audits, the service eliminates the complexity of implementing a data protection strategy for AWS customers. Clumio’s engineering team, led by Kumar’s co-founders Kaustubh Patil and Woon Ho Jung, has fully embraced ‘cloud-native,’ with an iteratively optimized microservices architecture and serverless computing services like AWS Lambda–or “Woon’s Army of Lambdas,” as the team refers to it–to parallelize these decomposed microservices, achieving exponential performance improvement with greater cost predictability.

Meticulous prioritization and focusing on the long-term growth opportunity

A key early product decision for the young startup was the careful prioritization of use-cases that would be relevant to the next generation of enterprise workloads. Pivoting away from an initial decision to address a broader set of use cases, Clumio chose to focus exclusively on data protection for cloud-based workloads, including native AWS workloads and VMware Cloud workloads hosted on AWS. Staying true to this vision has meant walking away from significant customer deals that didn’t align with their long-term growth objectives. But the opportunity cost of having a lean team focus on transient wins was deemed too high. “It’s not a question of if; it’s more a question of when [companies make the transition to Cloud and SaaS],” the CEO observes. Instead, Clumio chose to focus on early cloud adopters that value the ease of use and continuous innovation that is native to a SaaS delivery model.

Building on AWS has accelerated Clumio’s time to market and time to value. In Kumar’s words, “the delight of working on the [AWS] platform,” with its API-based, developer friendly interface and technological maturity, was a key factor in Clumio selecting AWS. As a pioneer in the space, AWS attracts and offers partners a valuable network of customers that are more amenable to adopting innovative technologies early. Clumio positions its offerings as complementary to the AWS platform and has effectively leveraged AWS’ broader ecosystem. It is fully integrated into an AWS customer’s existing end-to-end flow, so customers can use their existing AWS budget, enterprise agreements, and billing for Clumio services, which are available on the AWS Marketplace. As the CEO notes, “Fundamentally, we sell simplicity.”

Delivering value consistently through each stage of the customer journey

Kumar acknowledges that the journey to the cloud is akin to “crossing the chasm.” The team focuses on making customers comfortable with cloud adoption, emphasizing a comprehensive security posture and compliance certifications. Clumio has also now launched Clumio Discover, a cloud backup optimization engine that lets customers inspect their decentralized AWS environment across regions, accounts and workloads, to get actionable guidance to reduce AWS backup expenses and validate their compliance and security footprint. Additionally, Clumio Discover provides a cost comparison analysis and recommendations for lowering your backup costs with Clumio Protect.

Confident in their product differentiation and ability to deliver innovation rapidly, Clumio is betting on a product-led growth strategy and enabling customers to try their services risk-free with trial credits and consumption-based pricing. For the highest efficacy, release management and customer support must be highly efficient. Clumio achieves this through a SaaS delivery model: “The whole customer support problem becomes that much easier and that much more compelling in the cloud world, especially when you’re truly building and delivering a service,” says Kumar. There is only one version of the service running at any point of time, which makes it simpler to update the service seamlessly to deliver incremental innovation frequently. With real-time data and insight into the customer’s environment, Clumio can proactively identify and solve problems in a customer environment, sometimes before the customers may see it themselves, completely abstract away the environment backup and data protection functionality, and manage the recovery process transparently in case of a bad event.

Clumio maintains a close symbiotic collaboration with numerous AWS services teams, including Amazon Elastic Block Store (EBS), Amazon DynamoDB, and AWS Outposts, providing input on new features and APIs and influencing the product roadmap, all while leveraging this visibility to better shape their own architecture. On the go-to-market front, Clumio collaborates with segment specialists on the AWS Partner Network and AWS Marketplace teams, ensuring strong customer reach across the AWS customer network.

Clumio is currently available in United States and Canada. There is significant customer demand to expand internationally beyond these two regions, and spinning up a new region is simple from a product perspective. However, Clumio wants to be well-positioned with enhanced support services and customer success management that will delight their customers before expanding their footprint to a new region.

Clumio is a secure, scalable backup as a service that provides turnkey data protection against ransomware in AWS. Meet all your compliance needs, rapidly recover from any data loss, and get actionable insights to optimize your AWS backup spend.

About AWS SaaS Factory

AWS SaaS Factory helps organizations at any stage of the software-as-a-service (SaaS) journey. Whether looking to build new products, migrate existing applications, or optimize SaaS solutions on AWS, the AWS SaaS Factory Program can help. SaaS builders and operators are encouraged to reach out to their account representative to inquire about engagement models and to work with the AWS SaaS Factory team. Visit the AWS SaaS Factory Insights Hub to discover more technical and business content and best practices, and sign up to stay informed about the latest SaaS on AWS news, resources, and events.

Menten AI Leverages AWS, AI, and Quantum Computing to Create New and Improved Drugs

Mikey Tom — Mon, 16 Aug 2021 23:20:15 +0000

Menten AI created the world’s first protein designed on a quantum computer. The feat has huge implications for the world of drug discovery and design—and ultimately for all of us who may benefit from novel therapeutics to treat diseases. That prospect of real-world impact is exactly what propelled CEO Hans Melo and COO Tamás Gorbe to co-found Menten AI as they were finishing up their PhDs; Melo’s in Machine Learning and Computational Neuroscience, and Gorbe’s in Biocatalysis, Protein Engineering, and Medicinal Chemistry.

Tamás Gorbe, co-founder & COO

The startup’s drug-discovery technology focuses on designing peptides, or small protein-like molecules that bind to proteins of interest—AKA, “drug targets.” Peptides are medium-sized molecules, and they have the potential to bind to both intracellular and extracellular proteins. This flexibility allows peptides to “target basically all potential drug targets, including a niche section, which other drug types cannot,” explains Melo.

And while peptides offer “astronomical possibilities,” all of those possibilities have also traditionally necessitated astronomical investments in time and money. Traditional drug discovery approaches “synthesize or express hundreds of thousands—sometimes millions—of molecules” before testing them against a target. Only around 0.01 percent of those do anything at all, meaning that the “vast majority are just waste,” explains Melo.

This is where Menten AI has stepped in, merging machine learning and quantum computing methods to drastically reduce the time and cost needed to design peptides, while providing “access to a new chemical space that we didn’t have before,” says Melo. Until relatively recently, there was no computational technology that could design peptides accurately enough to actually attach to their drug targets.

Hans Melo, co-founder & CEO

By using computational modeling simulations, Menten AI can not only design new molecules from scratch for a specific target of interest, but also narrow down the options before bringing the molecules to life in the lab. “We do hundreds of thousands of in-silico simulations and designs,” explains Melo. “This is all on the computer, and then at the end, we can actually rank a small list of it, so we can pass 20 to 100 molecules and actually get very high success rates from those.”

Menten AI builds all of these simulations on AWS, using thousands of CPUs and GPUs. By doing the bulk of their simulation and testing in the cloud, the costs are drastically less than the previous methods, both from a time and money perspective.

“AWS has basically all the things we need from that perspective,” says Melo. “And even now, they’re building Amazon Braket, so there’s even more possibilities to do more of the quantum work that we’ve been developing as well.”

And it’s this quantum work that is increasingly important as Menten AI looks to the future: while machine learning has been used in drug discovery in the past, quantum computing is opening up entirely new approaches to solving hard problems.

One of those problems, combinatorial optimization, is at the heart of the type of peptide design that Menten AI conducts. Combinatorial optimization is “a search over a huge, huge, huge space,” Melo explains, “so you have trillions upon trillions upon trillions of possibilities.” It would take thousands of years to sort through each of these possibilities on a classical computer, and even algorithms that are designed to try to make this easier do not scale well, “meaning that as the problems get larger, you’re not able to actually find solutions to those problems.”

Quantum optimization algorithms, on the other hand, unlock immense parallelism, made possible by modeling many solutions simultaneously, allowing for exponential scaling. Menten AI has already had success in harnessing quantum computing for peptide design, and Melo is optimistic about continuing to use it in tandem with classical computing to discover solutions that were previously unavailable: “As quantum computers get better and bigger, we’re going to be able to leverage those technologies to improve the drugs that we’re creating.”

Next Generation Data Management for Clinical Trials & Research Built on AWS

AWS Editorial Team — Thu, 12 Aug 2021 16:54:34 +0000

Guest post by Steve Bilawey, Co-Founder & Chief Business Officer, Precision Digital Health

Data and Automation in Drug Development

The life sciences industry is absorbing a deluge of Real World Data (RWD). For the players who embrace technology and automation, this can revolutionize the clinical trials process, reducing time to market for life-changing treatments. Sponsors who can adapt their data management capabilities will see increased efficiencies across all phases of the clinical trial process, reducing drug development timelines, and generating more holistic patient views for better insights.

Cumbersome, disparate data sources in highly regulated environments have historically obstructed longitudinal patient views in clinical research. To help clients generate maximum evidence from trials, Precision Digital Health (PDH) developed a cloud-based platform capable of integrating and harmonizing disparate data assets in the R&D and life sciences industry. SUMMA is a fully compliant, next generation Platform-as-a-Service (PaaS) that enables data management for all aspects of the trials process, with self-service analytics and visualization tools to support next generation research needs. Built for rapid adoption and custom configuration, PDH’s technology innovations accelerate clinical evidence, resulting in faster patient access to better care.

Regulatory Compliant Data Management on SUMMA

SUMMA was designed and developed with patient and data security as its foundation to meet today’s regulatory and compliance requirements globally. It was purpose built with protections in place to separate between de-identified data and Protected Health Information (PHI). PDH has deep experience building validated solutions with sensitive clinical data in compliance with a stringent, global network of regulators (FDA, GDPR, HIPAA, GxP).

To provide an automatedlongitudinal visualization of a patient’s natural history, requires linking patient data across systems. To do so, PDH employs an FDA and HIPAA compliant, secure tokenization process. Tokenization impacts the entire clinical workflow from feasibility through the duration of the trial and prospectively.

Interoperability and AWS Integrations

A key aspect of SUMMA is its ability to integrate various AWS services, mobile devices, and industry clinical data feeds to enrich various analytics, facilitate insights, and generate evidence from trials. The SUMMA platform aggregates data from disparate sources and then standardizes it into a common data model (CDISC, OMOP) for harmonization and curation (eg. auto-cataloging, auto-fill, elastic search). This approach provides rapid, automated, real-time solutions across the organization and every aspect of the trials process, thereby reducing resources and time to value for clients. Data asset integrations include: Electronic Health Records (EHR); current and historical clinical trials data (EDC); pharmacy & medical claims data; diagnostic & specialty data (e.g. genomics); mobile eCOA/ePRO solutions; medical devices & wearables data.

PDH’s mobile solution facilitates a secure connection between clinicians and patients, regardless of geography. This expands research opportunities while improving data quality and patient experience. The direct-to-patient model flows into SUMMA as a single hub for all data including novel RWD endpoints such as sensor data, eCOA/ePRO and other patient centric metrics (e.g. food diaries).

Architecturally, SUMMA microservice framework provides an event driven data processing engine. SUMMA acts as an application layer that interoperates directly with key AWS components to provide highly customizable metadata driven microservices for various types of analytical and trial management workflows. These configurable data processing pipelines leverage a series of PDH custom connectors supporting key dataflows and AWS-based ingestion services. To adapt to any client size, SUMMA leverages containerized deployment using Kubernetes and Docker. This automated process using Terraform and Ansible allows us to deploy many different types of infrastructure configurations required to support various client needs. With this automation, clients can adjust the compute power they require in real-time without any downtime.

As visually represented above, the SUMMA platform is an industry specific layer within the tech stack that provides full life cycle data management for clinical data. SUMMA is built on underlying AWS Technology and easily integrates with a variety of AWS tools to enable self-service analytics for end-users:

· Amazon Comprehend – apply NLP to unstructured data sources

· Amazon SageMaker & Deep Learning – Machine Learning (ML)

· Amazon QuickSight – generate advanced reporting with visualizations and dashboards for multiple devices

PDH Powered by AWS

PDH’s SUMMA platform enables compliant evidence generation powered by AWS technology, with a configurable nature that keeps innovation at pace with the rapidly evolving healthcare and life sciences landscape. SUMMA is a repeatable, easily deployable PaaS that provides automated solutions for pharma and the broader healthcare industry. The microservices that underpin the platform enable rapid adoption and improvement of novel data sources and clinical workflows. SUMMA can work within existing client infrastructure or quickly bring new data lakes/capabilities to clients. SUMMA’s innovative technology automates, integrates and extends to new research capabilities. The built-in machine learning and AI capabilities turn real-world data into actionable insights.

Impact of Machine Learning & Advanced Analytics

SUMMA’s powerful and precise data capabilities allow for the use of advanced ML/AI techniques. To establish new patterns and insights for clinical discovery, SUMMA’s automation and linking capability enables broader impact across large volumes of data, in less time. SUMMA operationalizes ML/AI by automating and transforming the ML/AI into standard data flows that generate actionable predictive outputs for business users.

Case Study: Machine Learning for Treatment and Prevention in Osteoarthiritis

Using ML, PDH developed a model to identify predictive risk factors leading to early osteoarthritis diagnosis, accurate treatment, and prevention of knee replacement. The SUMMA platform prepared and transformed data to generate key clinical risk factors that would rapidly assess and risk stratify patients. The application of this model aided in early diagnosis and demonstrated a 90% accuracy rate in predicting osteoarthritis one year in advance. Additionally, SUMMA was able to automate the clinical workflow for targeting the patients accurately.

This discovery positively impacts multiple stakeholders through more efficient patient identification and treatment leading to decreased resource burden:

· Patient – physical benefits/quality of life, reduced spend and healthcare burden

· Health systems – early detection savings estimated at 25%

· Sponsor – improved patient targeting increases revenue 15-20%

About Precision Digital Health

PDH is a Platform-as-a-Service (PaaS) business, founded in 2015 to provide software development and solutions for clinical research and trials to clients in the Life Science/Healthcare industry. Core customers are Contract Research Organizations (CRO) and Pharmaceutical, Biotech and Research Organizations. PDH utilizes cloud-based open architecture (“interoperability”) with the most advanced technology and distributed computing to power SUMMA, a regulatory compliant, end-to-end data and analytics platform. Summa specializes in the use of Real World Data (RWD) to generate Real World Evidence (RWE) in support of clinical research, drug development, and improved health-related outcomes. Our platform is purpose-built with configurable data pipelines to preserve compliance across geographies and sectors. PDH services are available globally today, with support in over 30 countries.

To learn more about how PDH can power your clinical research, visit https://precisiondigitalhealth.com/ or send us a message at info@precisiondigitalhealth.com

Understanding and Optimizing ML Models with DarwinAI

Mikey Tom — Thu, 12 Aug 2021 16:34:23 +0000

Artificial intelligence may be the future, but 90% of AI models developed today don’t make it into production. “Suppose you’re able to get past the first part, where you manage to build something functional, and you manage to get through the second part, where it’s scalable,” says Dr. Alexander Wong, co-founder and Chief Scientist at DarwinAI. “Then you have no idea how it behaves in production, not only internally, but how it behaves in the wild. You don’t really know how to improve it, and you don’t really know where the underlying biases are.” This is known as the explainability issue.

When Wong set out to co-found DarwinAI, he was cognizant of those challenges. “I don’t like to build things that people will never use,” says Wong. “And one of my main goals and missions is to make sure that the types of research technologies that we build become tangible, tractable, impactful solutions to enterprises and society at large.” So he created DarwinAI, a technology program that centers around quantitative explainability and enables users to understand underlying systems of behavior. This also translates into increased ease when it comes to scalability. “Our underlying platform technology will actually accelerate and automate this whole process so that you end up with an AI model that meets your requirements, not just in terms of performance, but also in terms of your key operational KPIs, such as speed, size, memory requirement, and energy throughout,” says Wong.

Dr. Alexander Wong, co-founder and Chief Scientist

So far, DarwinAI has proven particularly useful when implemented by industries such as healthcare, manufacturing, and finance. With the help of AWS’s resources, DarwinAI was able to build a dependable, more efficient, state-of-the-art AI model for predicting lung function decline from pulmonary fibrosis. The explainable technology allowed developers to be aware of hidden data issues that might otherwise go unnoticed. “We perform what we call ‘explainability-driven data auditing,’” says Wong. “Using the explainable technology, we were able to surface things like imaging artifacts, artificial padding, dynamic range problems, and calibration issues. Those are all things that we can do from a data perspective—as well as model perspective—to ensure that the underlying AI solution is one that you can trust and depend on.”

When it comes to manufacturing, DarwinAI is similarly vital, especially for situations in which a human inspector needs to examine a huge number of parts, thus driving up the likelihood of human error. DarwinAI has been working with manufacturing partners to build solutions that minimize defects and problems that might result in error, as well as understand why these particular things are being flagged as problematic.

This explainability model has also been helpful in finance, an area where algorithmic transparency is especially important—both for building AI models and for client accountability. “We’re able to provide reasoning,” says Wong. “So a financial institute can report back to their client, ‘this happened because X, Y, and Z, and it’s critical for regulatory compliance.’” This sort of auditing can also be used to analyze the behavior of the AI models, searching for potential inherent demographic biases: “Using this type of quantitative explainability, we are not only able to surface how decisions are made, but ask: What are the key factors? Is it payment defaulting history? Is it because of gender? And so on and so forth.”

One of the most significant challenges for DarwinAI has been creating the technical tools—and deployment strategies—to help clients in dramatically different industries. This is where AWS has been particularly helpful. “AWS really helps when it comes to the flexibility we have in aligning with these different operational scenarios,” says Wong. “There are ones that require more computational resources because, for example, their manufacturing and inspection tasks are more complex. Or they have more plants, or they have more sensors. With AWS being so flexible and dynamic in serving these particular needs, it makes our life much easier to be able to customize different solutions for different partners.”

Looking forward, Wong believes that DarwinAI will be able to continue to serve an array of clients in even more diverse industries. “I think the key thing that makes us particularly unique is just how integrative our underlying technologies are to tailoring for specific needs,” says Wong. “The way our technology platform has been created allows for very tailored, unique solutions catered to very specific operational requirements. It’s not the one-size-fits-all, big-box-store approach; customization and tailor-made solutions is a very core part of what we’re enabling.”

How HeyJobs Ingests Millions of Jobs with AWS Lambda

AWS Editorial Team — Wed, 04 Aug 2021 01:54:09 +0000

Guest post by Gokay Kucuk, Engineering Manager, Inventory – Integrations Team, HeyJobs

At HeyJobs, we’re aiming to be the leading platform for those looking for the right job to live a fulfilling life. Serving millions of job seekers means that we need to ingest hundreds of thousands to future millions of job-offering details, multiple times a day, every day.

In this blog post, I will share our learnings about the AWS services we utilized for our serverless transformation. At the end of this transformation, the job ingestion capacity of HeyJobs grew from few hundred thousand to few millions per day while reducing our costs by ~30%.

In our first iteration of this ingestion service, we decided to make the inventory project its own service written in Ruby on Rails, as Rails was the most known framework in our company with PostgreSQL as the database. We utilized Sidekiq (a Rails job scheduling library) for background workers and data processing. The project was running on Amazon Elastic Container Service (Amazon ECS), and it was using RabbitMQ as the main communication tool with our monolith.

At first, our service was running smoothly, but with time, the number of job postings we processed grew from thousands a day to tens of thousands, and then to hundreds of thousands. This made us push more and more processing to the database layer with long and complicated SQL queries to be able to process everything in time. But it was still not enough. We still experienced high variance in processing times, which made us uncertain about our scalability metrics. Deploying larger instances would’ve solved the problem only temporarily while growing our costs by a large margin. We had to do something differently, but what?

Signal for Serverless Transformation

An important aspect of our project was that it needed to import job details at certain times, which created a very spiky usage pattern. We believe spiky usage is one of the most important triggers for serverless transformation. You can find the CPU utilization of our database server below.

Overview of Solution

Given that we needed to scale to accommodate spiking workloads, we decided to use Amazon Aurora Serverless since it scales to meet our demand. Combined with AWS Lambda functions for code execution, this setup gave us the greatest scaling possibilities, both up and down. We still wanted to keep our Rails monolith and minimize disruption, but our RabbitMQ provider had a hard cap on messages/sec processed. We decided to replace it with an Amazon Simple Queue Service (Amazon SQS) queue to remove this limitation as our processing was going to scale with demand.

We also decided to put another SQS queue at the start of our process to handle scaling windows of Aurora Serverless.

Amazon SQS and Amazon SNS

We utilize SQS and Amazon Simple Notification Service (Amazon SNS) heavily in our application. While we already had RabbitMQ queues for internal messaging, it was a rather alien component for most of the developers and did not provide any specific benefits for our use-case. Switching to SQS with Serverless Framework and deploying a new queue for a new function became easy and part of development. The more functionality we exposed with SQS as the trigger, the more we noticed that we ended up with more decoupled services we could build on top of.

Our application–based on event-driven architecture–has 2 main methods of communication:

Commands: We use SQS for triggering commands. We do care about what happens when a command fails within our domain, and having them in a queue lets us retry or route them. We try to make sure every command gets executed.
Events: We use SNS for publishing events. We don’t really care about what happens to an event, because it’s mostly going to be consumed from other domains/teams. We try to subscribe to this event from SQS queues from the domain that wants to consume the event. This creates a clear boundary between teams/domains.

Idempotent Operations

Using SQS as the main source of trigger for our Lambda functions required us to make our operations idempotent. If the SQS queue is not set-up in FIFO mode, it is an “at-least-once-delivery” queue system. This means sometimes it can deliver the same message more than once. While processing a low number of messages in SQS, we didn’t see any duplicated deliveries, but as soon as the number of messages in a queue grew to millions, we started seeing duplicate deliveries. This pushed us towards always trying to make our operations idempotent. As a bonus, making our operations idempotent made it easier to integrate other producers from our service landscape to the same SQS queues as we had to worry about states in our functions less.

Aurora Serverless

Aurora Serverless was the most novel component of the entire stack when we designed it, but after researching its capabilities we decided to use it as our main database. We had Lambdas as the main load generator to our database with SQS. Both components can scale fast. Having an ordinary database server would need us to either:

Deploy a database instance that ccouldan handle the peak load, but it would be wasteful as our loads were spiky.
Deploy a database instance that is rather small, which would cost less, but it would make our processing longer.

By choosing Aurora Serverless, we managed to both pay only as much as we needed while being able to scale up to very large instances to speed up our processing. We found out its scaling characteristics were a good fit for fast SQL queries, as having analytical long-running SQL queries makes it harder for Aurora Serverless to find scaling points.

Docker and Lambda

A recent update to Lambda made it possible to run docker images with Lambda as runtime. Previously, Lambda had a hard limit of 250 MB on the amount of source/library code that a Lambda function could have. The code also needed to be deployed in a special zip format, which most teams were not used to if they weren’t deploying AWS Lambda functions.

With the docker image support, the new size limitation for AWS Lambda is 10GB. This disk space is forgiving when it comes to loading large libraries and having big codebases. The image can be created from any CI system and published to an Amazon Elastic Container Registry (Amazon ECR) repository. This creates a perfect boundary for cross-team projects.

This method simply let us plug a docker image from another team into our processing pipeline, while having an infrastructure-as-code form Serverless Framework let us easily deploy a testing platform for other teams.

Conclusion

We saw lots of improvements during our move from a containerized Rails application to a serverless python application both in reduced complexity(~70%) and reduced costs(~%30). Our initial target for moving our application to a serverless architecture was mainly driven by our need to scale up and down many times during the day while trying to minimize development and running costs. While scaling could be also achieved with containers, it wouldn’t have been as easy as scaling Lambdas. We had to change our mindset and approach to problem-solving in certain situations. We started designing our applications in a more event-driven architecture with following domain driven design principles. This setup gave us a similar comfort level when it came to development, in comparison to the Rails environment that we were used to. After running our new application for a few months, we are happy with the results and have started to leverage the serverless setup in other areas. Follow us on our tech blogto learn more!

About the Author

Gokay Kucuk is an Engineering Manager working at HeyJobs Inventory and Integration teams, and he is always interested in learning more and utilizing the latest and greatest technology available. His main interests are serverless, blockchain (sauerbot), generative art (pollinations.ai), and photography. You can reach him on LinkedIn or gokay@hey.com

PulpoAR Uses Machine Learning to Build an Augmented Reality Shopping Experience for Beauty Products

Mikey Tom — Thu, 29 Jul 2021 22:45:27 +0000

Rayan Godoi was working inside a different startup when he and his future co-founders began to see the writing on the wall. In this case, the wall was real, but the writing was more like a computer-generated image superimposed on top of it: augmented reality, they realized, was the way of the future. Their work at the time involved connecting devices to the online world. “We learned how to build AR filters using augmented reality back in 2014,” Godoi recalls. “And we realized that that technology was really merging the physical and digital worlds together.”

Suddenly, the investments they saw industry giants like Google, Apple, and Facebook making in AR and VR technology began to make a lot more sense. The world was rapidly accelerating toward a digital-physical hybrid, and Godoi and his partners were keen to play a part in creating that (augmented) reality. “We want to accelerate the merging of these two worlds,” Godoi explains. “We had to start somewhere and we saw that, in the market for beauty, there’s real opportunity.”

As early as 2018, he recalls, the team realized, “One of the main applications that we could start working on immediately—because companies were already searching for these—was the virtual try-on of products using augmented reality.”

Rayan Godoi, Cofounder & CRO

Soon after that, PulpoAR was born. “The shopping journey is broken,” Godoi explains. E-commerce accounted for nearly 20 percent of retail sales worldwide last year—the largest share ever—but when consumers shop online, they can’t try new products on, and that’s a barrier that keeps many shoppers, particularly in the beauty industry, from clicking the check-out button. That’s where PulpoAR comes in: “We create augmented shopping solutions that allow customers to try on makeup looks anywhere, anytime they want, with precision and realism.”

To build a service that could faithfully recreate the experience of trying on a beauty product in-store, the PulpoAR team had to gather countless images, then teach their machine-learning model to analyze and programmatically identify them. “We were using different platforms,” CTO Bugrahan Bayat recalls. “We were training our model in different ways. We had people tagging the pictures and teaching it to watch and recognize the pictures. But we realized there was another way to do this—and then we started using the tools available from AWS.”

Today, the team runs its ML models using Amazon SageMaker and AWS Lambda; SageMaker trains the algorithm in face detection, segmentation, and image processing, while Lambda acts as the service’s serverless computing platform to enable seamless production and scale.

PulpoAR performs 100 percent of its operations with AWS. Photos taken by users are uploaded and analyzed with Lambda functions before being sent back to the customer’s browser. “Cloud solutions are very important for us because our technology requires serious device processing power,” Bayat explains. “With AWS, we were able to access the same processing power on every device, and thus our users achieved better results.”

The serverless processing power and reliability of AWS’s tools were key to PulpoAR’s growth. Without AWS, Godoi says, “It wouldn’t be possible to scale at all. We had to run millions of pictures to get good-quality computer vision. This is only possible with AWS features, that’s for sure.”

Buğrahan Bayat, Cofounder & CTO

Today, the company—via partners like Sephora, MAC Cosmetics, and Flormar—is processing two million try-ons every month. “We want to raise it to 60 million try-ons per month,” Bayat says. “We can only achieve that with a scalable architecture on the back end.”

In the meantime, they’re considering ways to marry their technology with other innovations, like the recommendation engine based on behavior. “We can say that if someone has black hair, a wide chin, and blue eyes, and that person tries this lipstick on for more than 30 seconds, then let’s suggest this eyeliner to her or him. But if the person does not have blue eyes or black hair, or if they have a narrow chin, it suggests a different product,” PulpoAR’s Head of Growth, Huseyin Oguz, explains. “If we can combine the biometric data and behavioral data to recommend a new product, I believe it will be a really interesting development in the near future.”

The company is eyeing an expansion into skincare analysis as well. “We want to use augmented reality to see wrinkles and skin marks and make suggestions of behavior or products or services for people to enhance their skin health,” Godoi explains.

All of this is, in Godoi’s mind, still just the beginning, both for the company and also for the rest of us. “We decided to launch our first product—the virtual try-on of makeup—then expand, generate revenue, and keep building new products to accelerate the merging of the physical and digital worlds.” With all of the possible applications—from healthcare and combat training, to tourism and immersive entertainment—it’s no wonder that Godoi and his colleagues are confident that, a decade from now, augmented reality and virtual reality, “Will be ubiquitous. It will be omnipresent. Everyone is going to be using these kinds of solutions.”

HackerEarth Scales Up Continuous Integration for Future Needs with AWS

AWS Editorial Team — Thu, 29 Jul 2021 21:02:19 +0000

Guest post by Payal Moondra, Staff SDET at HackerEarth

HackerEarth provides enterprise software that helps organizations with their technical hiring needs. HackerEarth is used by organizations for technical skill assessment and remote video interviewing. Since its inception in 2012, HackerEarth has built a base of 5.5M+ developers, along with a community around it.

One of the goals of every fast-paced organization is to have a Continuous Integration (CI) pipeline that ensures every check-in is best verified before it can be pushed to production. HackerEarth wanted to achieve a CI model that has enough safety nets for every check-in that goes into each Pull Request (PR), as well as make the process scalable and cost effective. The safety nets in the CI pipeline includes unit tests, integration tests, functional tests, security tests, and static code analysis. These safety nets in the pipeline provide constructive feedback for the PR, and the necessary steps are then taken to mitigate the gaps. For integration tests in this pipeline, HackerEarth used AWS CodeBuild along with Amazon S3 and Amazon Elastic Container Registry (Amazon ECR).

What does Continuous Integration mean?

For a robust CI pipeline, each line of code written should be well tested before code changes are merged back to trunk in the code repository. For every check-in, HackerEarth wanted to have a pipeline that runs a series of tests (static code analysis, unit, integration, security tests among others) to ensure its mergeability (pass a set of checks before they can be merged). This helps in identifying major issues well in advance in the deployment cycle and prevents issues from going to production. While HackerEarth had this setup in place for some time, it had a few disadvantages, including high cost, high maintenance, and scalability limitations. The team was looking to solve this problem and prepare it for future scalability needs. In the process, HackerEarth discovered AWS CodeBuild, a fully managed continuous integration service that compiles source code, runs tests, and produces software packages that are ready to deploy. Before we jump into how AWS CodeBuild has simplified the process, let us walk through the problems associated with our former architecture.

Former Architecture:

Here is what would happen in the former architecture when a PR was created or updated by a developer:

1. Jenkins webhook would trigger Integration tests.
2. The pipeline would divide Integration tests into two parts and send them to two Amazon EC2 instances. The tests would run in parallel in these EC2 instances to reduce the total execution time.
EC2 Instance 1 : NODE_TOTAL=2 NODE_INDEX=1 nosetests --with-parallel
EC2 Instance 2 : NODE_TOTAL=2 NODE_INDEX=2 nosetests --with-parallel

Note: These two EC2 instances that ran all the time would have already been configured with required services like Redis, MongoDB, rabbitMQ, and MySQL database.

3. After the tests, Jenkins would collect the test reports from both machines and merge them into the final HTML report.

Problems with the former architecture:

High Maintenance:
EC2 instances required constant maintenance. This included making sure required services are always up and running, stopping EC2 instances when they aren’t being used, periodic volume cleanup for maximum space utilization, keeping the directory counts in check etc.
Cost for unused resources:
The need to keep the EC2 instances up and running 24×7 incurred costs even on days when there weren’t many PRs to make use of them all the time.
Scaling:
In order to reduce queue time for PRs, they need to run in parallel. The issue with the existing setup is that every PR needs to have its own code directory. Hence, only a limited number of PRs could run at a time, making the pipeline less scalable.
Multiple concurrent builds running on common set of instances:
If something went wrong with either of the two instances, it affected all the tests on them.
One database shared across multiple builds:
Since we had our instances set up in advance, every build would use the same database, and that created conflicts when the nature of PRs was different.
Longer wait time:
Every PR needs to check out its own codebase directory. Even with a larger EC2 machine, tests for not more than two PRs could run at a time (without bumping up the size of attached EBS volume). So every PR triggered after the first two would be in the queue and that delayed the feedback for them. This led to a situation where continuous integration wasn’t happening quickly enough, and the feedback was delayed.

HackerEarth’s goal was to fix all the problems mentioned above to simplify the CI pipeline and make it as efficient as possible. The team realized AWS CodeBuild was the single answer to all the problems. It is a fully managed continuous integration service that compiles source code, runs tests, and produces software packages that are ready to deploy. With CodeBuild, you don’t need to provision, manage, and scale your own build servers. It scales based on your build needs (observing service limits), and you pay for only the time you use it!

The improved, simplified, and cost optimized CI architecture:

Here is what would happen in the new architecture when a PR was created or updated by a developer:

Jenkins webhook triggers AWS CodeBuild to read buildspec.yml file stored in the project directory.
The buildspec.yml (from the project files) has details of docker images stored in Amazon ECR. Amazon Codebuild pulls the image and launches.
AWS CodeBuild project can run a number of parallel builds corresponding to the scenarios defined in the Jenkins pipeline.
Once the execution within the container completes, including persisting the output XML file into S3, CodeBuild destroys the container.
Jenkins creates the final HTML report by combining the XML files in S3 (using jenkins junit plugin).

Since tight integration with Jenkins already existed, the team leveraged AWS CodeBuild Jenkins plugin (instead of AWS Codebuild support for webhooks with Bitbucket).

Creating an AWS CodeBuild Project:

1. Source:
Source provider: Bitbucket
Repository URL: https://bitbucket.org/<YOUR_USER>/<YOUR_REPO>.git
Source: Leave it blank because will be managed by branch name environment variable from Jenkins.

2. Environment:
Environment image: Custom image
Environment type: Linux
Privilege: Yes

3. Buildspec:
You can create a file called buildspec.yml in your project directory. A sample of builspec.yml shown below pulls an Amazon ECR image and runs its own container, sets it up with required services and DB (setup.sh), runs the tests, and uploads the results to S3 for each build.

4. Artifacts:
Amazon S3 (Create a bucket in S3 and provide the bucket name in the configuration. Also, provide a namespace as BUILD ID if you want the artifacts for each build separately.)

Sample Jenkins Codebuild plugin configuration:

def setUpCodeBuildEnv(PR-id){
//Builds the image and push it to ECR tagging PR-id
}

def executeTestAndGenerateReport(tests, total_nodes=2, node_number=1){
if (env.BRANCH_NAME.startsWith('PR')) {
   branch = "${CHANGE_BRANCH}"
}
else {
   branch = "${env.BRANCH_NAME}"
}

env_variables = "[{TESTS, " + tests + "}, {NOSE_NODES, " + total_nodes + "}, {NODE_NUMBER, " + node_number + "}]"

awsCodeBuild credentialsId: 'codebuild_creds',
         credentialsType: 'jenkins',
         sourceControlType: 'project',
         envVariables: env_variables,
         sourceVersion: branch,
         projectName: 'codebuild_ci',
         region: 'region_name',
         downloadArtifacts: 'TRUE'
}

Sample Jenkins pipeline configuration:

pipeline {
	stages{
		stage(‘Divide tests’){
		tests_1 = divideTests(3, 1)
        tests_2 = divideTests(3, 2)
        tests_3 = divideTests(3, 3)
		}
		stage(‘Parallel CodeBuild - 1’){
			executeTestAndGenerateReport(tests_1, 3, 1)
        }	
		stage(‘Parallel CodeBuild - 2’){
			executeTestAndGenerateReport(tests_2, 3, 2)
        }	
	  	stage(‘Parallel CodeBuild - 3’){
			executeTestAndGenerateReport(tests_3, 3, 3)
        }
    }
    post{
	    //merge all artifacts received from parallel CodeBuild builds above using jenkins junit plugin
    }
}

Sample buildspec.yml file:

Refer to the build specification reference for CodeBuild for additional details.

version: 0.2
env:
 variables:
     NOSE_NODES: 2	 //passed from Jenkins codebuild plugin, this is used for dividing tests using django nose-parallel plugin
     NODE_NUMBER: 1	//passed from Jenkins codebuild plugin, this is used for dividing tests using django nose-parallel plugin
     TESTS: ‘tests’		//passed from Jenkins codebuild plugin, this variable container test name in case if we want to run an individual test instead of all. Otherwise left blank
     ECR_IMAGE: ‘image_uri’	//Base image stored in AWS ECR. It is the image that is built and pushed to ECR in setUpCodeBuildEnv build for each PR
 parameter-store:
     ACCESS_KEY_ID: /CodeBuild/ACCESS_KEY_ID		//created inside AWS parameters store
     SECRET_ACCESS_KEY: /CodeBuild/SECRET_ACCESS_KEY	//created inside AWS parameters store
phases:
 install:
   runtime-versions:
     docker: 18
 pre_build:
   commands:
     - aws configure set aws_access_key_id $ACCESS_KEY_ID
     - aws configure set aws_secret_access_key $SECRET_ACCESS_KEY
     - eval $(aws ecr get-login --no-include-email)
     - docker pull '$ECR_IMAGE' || true
     - docker run --name test_container -u root -t -d --volume $CODEBUILD_SRC_DIR:<project_dir_path> tag_name:latest
     - python update_db.py download_db	//Script to download  base DB schema stored in Amazon S3
     - sh setup.sh	//created inside project root to run services and create DB
 build:
   commands:
     - docker exec -u root -i -e node_conf="$TESTS $NODE_NUMBER $NOSE_NODES" test_container bash -c ' python code_build_run.py ${node_conf}'
artifacts:
 files:
   - test_report/tests_report_$NOSE_NODE_NUMBER.xml

Conclusion

HackerEarth solved all the issues in their former architecture and reduced 50% of the cost per PR build using AWS CodeBuild, Amazon S3, and Amazon ECR. The benefits included:

Zero Maintenance:
There is absolutely zero maintenance with using AWS CodeBuild. We do not need to maintain any machines at all.
Cost Efficacy:
The pricing structure for AWS CodeBuild is per minute. Hence, unlike the old architecture where there was a fixed cost for keeping EC2 machines up even when not used, we only pay for the duration the service is in use. We have achieved 50% reduction in cost per PR build with AWS CodeBuild.
Continuous Scaling:
As checking out source code became part of AWS CodeBuild build, we didn’t need storage space for Jenkins node, so we could increase the number of executors to 15, which meant 15 PRs would run in parallel without zero wait time.
Independent runtime environment:
AWS CodeBuild provides separate container and allows for source code specific to each build. Each build runs on a completely isolated environment (docker container). This means if something goes wrong in one build, it doesn’t affect any other builds.
Zero Interdependency:
We could now set up DB for each PR individually in AWS CodeBuild. This meant that multiple builds do not share database resources. Each PR will run on its own database and decommission it once done.
Zero downtime:
We don’t have to worry about something going wrong in pre-setup build environments. AWS CodeBuild takes care of that undifferentiated operational effort.
Faster Build times:
With increased parallelism within each PR, the overall build time decreases in proportion to the number of builds running in parallel. For example, the PR running on 2 EC2 instances takes around 30 mins. This same PR would take 15 mins if we run them on 10 parallel AWS CodeBuild builds (While keeping the cost in check!)

All in all, AWS CodeBuild helped HackerEarth reduce cost, reduce build time, all while taking care of managing the undifferentiated heavy lifting of maintenance.

FutureFit AI Is Building a GPS for Your Career

Mikey Tom — Wed, 28 Jul 2021 23:05:29 +0000

Hamoon Ekhtiari was only in high school when his family emigrated to North America, but he still remembers the pressure he felt to forge a path in his unfamiliar new home. “Very early on, I started to have to figure out—as someone who didn’t know the country, the landscape, or the market—what jobs or careers I might want to pursue,” Ekhtiari recalls. Watching his parents, immigrant professionals who experienced their own struggles having to change their names or accept positions below their level of education, Ekhtiari knew it wouldn’t be easy.

He went on to work in human resources and later consulting, but the memory of those early days stayed with Ekhtiari. He wanted to change the way that people navigate their career transitions, help people like his parents—like his younger self—chart a course for themselves under difficult circumstances.

Hamoon Ekhtiari, Founder & CEO

“Increasingly, the future of work is changing in a direction where pathways to opportunity and prosperity are becoming less and less visible, predictable, and manageable by individuals than in the past,” Ekhtiari says. That led him to team up with co-founders Terralynn Forsyth and Zeeshan Shahid to create FutureFit AI, which partners with enterprises and governments to accelerate workforce transformation by providing individuals with an AI-powered tool that they like to think of as a “GPS for your career.”

The trio learned quickly that there was significant demand for the product they had envisioned. “We spoke to several hundred people who had gone through layoffs, and the three most common words we heard were: I am ‘ashamed,’ ‘lost,’ and ‘confused,’” Ekhtiari says. Those feelings of “being shattered and crushed and then not knowing where to go next” became the founders’ starting point.

FutureFit AI’s platform is designed to help guide individuals through this process, Ekhtiari explains, by first figuring out where they are, then gaining an understanding of where they want to go and helping them map a course between points A and B. The tool starts with assessments: what qualities, skills, and experiences does a jobseeker already possess? Then, FutureFit AI’s algorithms generate a list of recommendations based on three criteria: what’s feasible for this person, what’s most desirable to them, and what will be sustainable over time.

“Sustainability is about the match of the career to the market, its potential for growth, and its risk of automation,” Ekhtiari explains. “There are questions of fairness and equity…we’ve had to architect our algorithms to consider things like: just because I am like somebody else, that doesn’t mean my recommendations should be only what others have done in the past, because job markets haven’t been as open—they have been more unfair—than they will hopefully be in the future.”
FutureFit AI started off as a relatively small team, but from the very beginning, Ekhtiari and his colleagues knew they wanted the ability to scale globally. That knowledge made AWS a natural choice. Before he came to the company, FutureFit AI’s Director of Engineering, Jon Morrow, spent almost a decade working in the infrastructure automation space, on cloud platforms like AWS. “I, too, have gone through a job transition where I was laid off from my position,” Morrow says, “And helping build FutureFit AI’s mission resonated with me deeply given that past experience.”

“When you look at the various clouds, AWS is the [most] established, well-understood one, and it made real sense for us to partner within the AWS environment to build out,” Morrow says. Thanks to its capability, the company expects to be live in a dozen countries by the end of the year. “And that couldn’t be done without an infrastructure support partner like AWS,” Ekhtiari says.

Morrow agrees. FutureFit AI just launched in the European Union and, he says, “Being able to leverage all of the automation that we’ve built and tested in our North America data centers to deploy to another country has really enabled the business to go out and sell where they want to sell. And from an engineering perspective, we’re able to support that without a lot of extra overhead.”

There is some concern these days about automation, and the havoc it’s poised to wreak across industries in the coming years, but Ekhtiari sees a potential to harness that power for good.

“The rate at which work is changing and what that means in terms of impact on people, especially those most at the margin, is significant,” he says. “AI, by its nature, is going to have those impacts, but it also has the power to be part of the solution: mitigating impacts by getting people to actually be able to take on work that fits who they are better, work that is more meaningful, that’s more enjoyable. But to do that, we don’t need to be protecting jobs. We do need to be protecting people.”

Brand Tracking with Bayesian Statistics and AWS Batch

AWS Editorial Team — Tue, 27 Jul 2021 19:12:12 +0000

Guest post by Corrie Bartelheimer, Senior Data Scientist, Latana

At Latana, we help customers make better marketing decisions about their brand by using advanced algorithms to track consumer perception. For example, a common problem with brand tracking is bias; while online surveys are relatively cheap, the respondents are usually not representative of the general population and the results can often falsely favor certain groups. Additionally, clients are often interested in how their brand fares in specific target groups, like households with young children, and very small target groups can be especially tricky to get reliable estimates from. The smaller the target group, the more difficult it is to distinguish between signal and noise: if there is a small change from one month to the other, is that due to random changes in survey respondents or is it actually a meaningful change? This last point is especially relevant for smaller companies that are not yet widely known, where a small absolute change implies a big relative change.

In this post, we outline how mathematical models and probability theory, specifically Bayesian methods, address some of the big problems in brand marketing and how AWS Batch, together with Metaflow, solves many of the technical issues that used to be major obstacles to using Bayesian methods at scale. Methods to compute Bayesian models such as Markov Chain Monte Carlo (MCMC) methods are infamous for being slow and computationally demanding, which is why adoption has been relatively low in the data industry in comparison to machine learning (ML) methods. Using Metaflow with AWS Batch allows us to provide a brand tracking solution powered by Bayesian statistics to our clients at scale.

Better Target Group Estimates through Hierarchies

At the core of our tracking tool is a statistical technique called MRP (Multilevel Regression with Poststratification), sometimes also affectionately known as Mister P. In its simplest form, MRP is a hierarchical linear model together with a poststratification. If we, for example, want to estimate how many people in the population know the brand FancyTech, and we are interested in the demographic’s gender, education, and age, we would set up the following model:

The variable `knows_brand` would be a binary variable (yes or no), so we use the inverse logit as link function. Following a hierarchical model, each demographic is grouped. This means that, for example, for education, we estimate one coefficient for each education level (low, medium, high), and all three coefficients come from a common probability distribution like Normal(0,1). This ensures that we get different estimates for each level but also helps regularize the estimates. If there are only a handful of respondents with high education, it’s better to stick closer to the estimate of the other two levels; otherwise, we risk skewing the data. Additionally, since we’re using a Bayesian approach, we get uncertainty estimates for free, which is especially important for small target groups where the estimate might be based on very few observations.

So far, we’ve addressed the problem of small target groups (through the hierarchical groups) as well as the problem of distinguishing between signal and noise (through the regularizing features of the Bayesian model). But how do we solve the problem of biased samples?

Poststratification of Biased Samples

The good news is that we generally do have information about how common different target groups are in the general population. Many countries collect extensive census data that give exact counts, like how many women live in Germany with a high education level between 40 and 50 years of age. If we have the proportions for each target group, we can use these to weight our predictions that we get from the hierarchical model. For example, if we only care about the gender demographics, we’d compute the prediction for both men and women and get the estimate for the general population as follows:

This formula easily extends to more demographics. As we’re using a generative Bayesian model, all uncertainty estimates also propagate across these calculations.

Challenges when Using Bayes

While using a Bayesian approach has multiple advantages, it also brings its challenges. Bayesian methods are known to be much slower than their frequentist counterparts and traditionally, their speed and computational complexity has been a major obstacle to using these methods at scale. Fortunately, recent developments such as the NUTS sampler and new probabilistic programming languages such as Stan mean that our models already run relatively fast; one of our models takes on average less than twenty minutes on a normal work laptop.

However, we need to run one model per question, and, depending on how the survey questions are coded, we also need to run multiple models per question. A question where a user can select multiple answers requires binarization, meaning we end up with one model per answer option. This also implies that one survey can result in a few hundred models, which would take a total compute time of multiple days. To speed up runtime, we therefore decided to run the models in parallel using AWS Batch, a service that simplifies running hundreds of parallel jobs in parallel. It was just what we needed.

Parallelizing Hundreds of Models

In each model, the same predictor variables are used (the demographic variables), and only the dependent variable changes (the question we predict). It thus makes sense to do all the data transformation once and only parallelize the models itself. Using AWS Batch, this means we would have one batch job to load and transform the data, then one job per model in parallelm and a single job afterwards to collect and combine the results. However, this also means we need to take care of data sharing between the different batch jobs. We quickly noticed that orchestrating the different batch jobs ourselves was difficult to maintain and was also taking away our time and resources that we’d rather spend on improving the statistical models.

In a previous project, our Data Science team had already worked with the framework Metaflow, and we concluded that it was also a good fit for our model pipeline. Metaflow structures a model pipeline, also called Flow, into multiple steps, and each step is computed in its own separate conda environment. We decided to create a few intermediate steps because it not only helps us to structure our code better, but it also gives us more fine grained control of the environment for each step and the resources needed.

To run the models in parallel, we create a single model step in our MRPFlow code and tell Metaflow to run it for each of the different question parameters. Metaflow takes care of saving and sharing data across the steps as well as handling the parallelization. One of the advantages when working with Metaflow is that there is barely any difference between running the code locally and running the code on AWS Batch. In our first outline of the model pipeline, we still had a few lines in our code that checked if it was running locally or on AWS Batch and then ran different code branches to either parallelize the models (on AWS Batch) or not (running locally). This made debugging the code rather difficult, and the process was generally more error prone. Many times, errors happening on AWS Batch were not reproducible locally. Using Metaflow ensures that the same piece of code is run, locally and on AWS Batch, and it also creates the same conda environment for both environments. This helps tremendously with reproducibility and generally makes development easier since we can be sure that what “works on my machine” will also work in the cloud and give the same results.

Results

By using a Bayesian approach, we can address some of the most common problems in brand tracking, and results even for small target groups are more reliable and less susceptible to random noise. For us, the benefits of Bayesian methods far outweigh the technical challenges that come with it. Furthermore, with Metaflow, we found a framework that takes care of those challenges for us and makes it feasible to run Bayesian models such as MRP at scale on AWS without requiring in-depth knowledge of AWS operations. Switching to Metaflow with AWS Batch really allowed us to speed up our development process and focus on improving the statistical models. In a previous version of the model pipeline, we ran everything on a single instance and could thus only parallelize a few dozen models. Moving to a fully parallelized setup meant we could run more than ten times faster and, on top of that, also have more flexibility in designing our models to deliver more accurate results as wells as smaller credibility intervals. The speed up in development has been especially noticeable in a recent project, where we had been working on stacking multiple models for each question, all run in parallel. Adding this to our existing MRPFlow was remarkably straightforward. In this sense, Metaflow has been extremely helpful in letting us focus on further improving our models to become more flexible and delivering more accurate results.

Creating Customized Business Applications with Clappia’s No Code Platform

AWS Editorial Team — Fri, 23 Jul 2021 21:45:32 +0000

September 8, 2021: Amazon Elasticsearch Service has been renamed to Amazon OpenSearch Service. See details.

Guest post by Ashutosh Kumar, Co-Founder and CEO, Sarthak Jain, Co-Founder and CTO, Clappia

Clappia is a no code platform where creating customized business applications is as easy as working with Excel sheets. Apps built on Clappia range from elementary to very complex, involving master data, automation workflows, and integrations with external systems.

The Business Problem

Businesses need applications to manage their processes. But procuring these applications can be expensive and time consuming. Some businesses hire developers to build the applications and set up an IT team to manage the life cycle of the applications. Others outsource the application development to freelancers and agencies. There are a few problems with these approaches:

● Both approaches consume time and money and are potentially unreliable.
● Business users need to decide where to host the applications, how to publish the apps on app stores, etc.
● Once applications are developed, any change in the functionality takes time to develop and release.
● Scale is not generally not considered during the initial phases, which leads to increased costs in the long term.

To overcome these problems, we created a platform where business users can create apps by themselves without the knowledge of any coding and without worrying about deployment, hosting, storage, change management, and scaling.

Our Approach

Business users come to Clappia and design apps per their understanding of the processes. They are presented with simple drag and drop blocks to design their apps, which consist of basic data input blocks to capture text, gps location, images, and barcodes as well as advanced blocks to consume data from other Clappia apps or external APIs. They can also configure multiple workflows like sending emails and sms, hierarchical approvals, sending data to external APIs. These apps are distributed to end users who submit data via the app. Upon submission, all the workflows are triggered and the admins get to see the submissions by all the users.

Our Challenges

The definitions of our apps can be very dynamic and very different from one another. App admins can change the definitions at any time, and the changes need to propagate to all the end-users immediately.
Users can distribute each app to thousands of users, and the number of submissions made per day by one user can range from 1 to 50,000, which means that the system should be capable of handling varying amounts of load in terms of requests per second.
There can be a lot of asynchronous processing required once a user submits an app, and the processing logic can be very dynamic. For simple use cases, it might involve sending emails or SMS. It might include fetching data from or pushing data to other Clappia apps and making external API calls for complex use cases.
Users of the app need real-time analysis of their data interactively.

Why Serverless?

Starting with a few Amazon EC2 servers was a quick option for us to go live. Given the complexity of apps and the amount of data we had to handle, we knew that the servers would become unmanageable very soon. We wanted to invest our time in building something scalable rather than being burdened by operational overhead. So we decided to go with the serverless model of deployment using various products that AWS offers.

Which AWS products helped us address these challenges and go serverless?

AWS Lambda: We developed multiple microservices as part of our architecture. Lambda makes our life easy; push your code, and one is sorted. After that, we don’t need to worry about managing servers, up-scaling or down- scaling servers. AWS Lambda takes care of all that.
API Gateway: All the business logic is exposed to clients using API Gateway, which serves as a proxy to the Lambda functions. It acts as a layer to provide RESTful API functionality.
Amazon DynamoDB with streams: DynamoDB Streams capture time-ordered sequence of item-level modifications in a DynamoDB table and durably store the information for up to 24 hours. Applications can access a series of stream records, which contain an item change, from a DynamoDB stream in near real time.
NoSQL database to store the definitions of our apps and workflows, as these definitions are very dynamic. Also, as our submissions data grew from a few hundred per month to a million per month, DynamoDB got auto-scaled to handle the load.
AWS Step Functions: Step Functions allow us to orchestrate multiple small business functions like sending emails, SMS, and making external API calls.
SNS/EventBridge: Amazon SNS provides pub/sub functionality provides topics for high-throughput, push-based, many-to-many messaging between distributed systems, microservices, and event-driven serverless applications. Another service for application integration is Amazon EventBridge. It connects applications using events. An event is a signal that a system’s state has changed, like a change in the status of a customer support ticket. Customers can integrate their own AWS applications with microservices. We do a lot of asynchronous processing on our data, publishing events to AWS SNS and EventBridge and then handling these events independently. This ensures that we don’t impact the latencies of our core functions.
AWS OpenSearch Service (successor to Amazon Elasticsearch Service): Amazon OpenSearch Service lets you deploy your Amazon OpenSearch Service cluster in minutes to provide a fast, personalized search experience for applications, websites, and data lake catalogs, and to quickly find relevant data.

We send our app submissions data to ElasticSearch from where it is queried to get summary and aggregations. Following is the high-level architectural diagram of our system.

There are four major components of the architecture.

1. App Definitions: Admins of the end consumer organization are presented with an app designer to create apps for their users. The designer is a single page web application developed on Angular, hosted on Amazon s3 and served through Amazon CloudFront. It offers simple drag and drop blocks for the admins to design the apps. The calls to save the app definitions are sent to the Definitions Microservice Lambda Functions via API Gateway, and these definitions are stored in a DynamoDB Table. This DynamoDB Table has Streams enabled, which triggers a Lambda that maintains versioning of these apps in another DynamoDB table.
2. App Submissions: Users of the apps access their apps on the web or mobile. The data submitted by these users come to the Submissions Microservice Lambda Function, again via API Gateway and stored in DynamoDB. In some cases, users upload images, videos and documents as well. These are stored in Amazon S3.
3. Workflows: Once the data is submitted and stored, asynchronous events are published on SNS and EventBridge, picked up by the Workflows Execution Microservice and sent to AWS Step Functions. Step Functions perform the tasks for sending notifications, making external API calls, running schedules/loops etc.
4. Analytics Engine: All the data is also sent to ElasticSearch Service. The Analytics Microservice queries ElasticSearch to perform aggregations on the data and present the admin users with simple interactive charts and data visualisations on a custom built application.

What benefits do we get by using the Serverless products from AWS?

1. Cost

By going serverless, cost is the most obvious advantage in the early days of the product. We pay only for the requests that we serve. This means that we need to pay almost nothing when we have a small customer base. Additionally, AWS Lambda gives 1 million free executions every month forever, which is a significant cost advantage for a B2B product. This helped us save a lot of money in the initial few months, and we spent it on our product development.

2. Serverless

If we manage our servers, too many operational tasks come up – monitoring CPU/memory utilization, restarting processes in case of any issues, fixing OutOfMemory exceptions, fixing networking issues, etc. This creates a 20-40% operational overhead on the developers. With the serverless model, we seldom have to bother with these issues. Developers can focus only on developing the core business logic.

3. Scale

The serverless model is inherently scalable. It takes zero hours of effort to scale a serverless application up or down. If customer requests went up from a few hundred per day to thousands or even millions per day, the serverless model would take care of provisioning infrastructure for application.

Conclusion

Overall, going with the serverless model helped us build and release our platform much faster. It is continuously proving to be beneficial in terms of maintenance, speed of development. and ease of scaling. We shall continue to leverage services from AWS platform to build our future applications and continue to publish on Clappia marketplace. As we are currently targeting more and more enterprise customers, we are expecting the number of users and the amount of data and users to grow by more than 5,000% over the next year, and we are confident that with the serverless model, we’ll be able to easily handle the increased amount of load.

SaaS Founder Series: Dremio’s Journey to Unicorn Status

AWS Editorial Team — Wed, 21 Jul 2021 18:40:37 +0000

Guest post by Afza Wajid, Bill Tarr, and Mark Birch, AWS

With the proliferation of systems of engagement and intelligence, there is an increasing emphasis on delivering data insights throughout modern organizations. However, in order for data to be analyzed effectively, specialized skills and expensive custom code are required to centralize data across disparate storage systems before it can be analyzed. With the advent of open-source distributed data storage frameworks like Hadoop, developers were able to directly query distributed data sources, but data scientists and analysts still could not derive value from data in a self-service manner.

Founder & CPO, Tomer Shiran

Dremio, a Series D funded unicorn startup founded in 2015, simplifies the analytics stack with a high-performance, highly-efficient query engine that enables data consumers to directly query petabyte-scale cloud data lake storage, while eliminating massive data transfers and vendor lock-in, and mitigating security risks. Additionally, a semantic self-service layer that connects data sources to intelligence tools improves time to value for data scientists and analysts. Dremio is now extending the reach of its solution with the launch of Dremio Cloud, a cloud-native data lake as a service that simplifies the customer experience.

The AWS SaaS Factory team invited Founder and Chief Product Officer of Dremio, Tomer Shiran, to discuss Dremio’s journey to software-as-a-service and to share key learnings for businesses building SaaS and platform-as-a-service (PaaS) offerings on AWS. Shiran, an entrepreneur with over 15 years experience in enterprise software, has held positions in product management and engineering at Hewlett Packard, Microsoft, and IBM Research. Prior to Dremio, he was VP of Product at MapR and helped grow the company from five employees to nearly 400 employees and hundreds of enterprise customers. Read on to learn more about Dremio’s journey to unicorn status.

SaaS Factory: Tomer, thank you for taking the time to speak with us today. To start, please tell us a bit about the value proposition on which Dremio was founded.

Tomer Shiran: Everyone wants more data. But the more data there is, the harder it is to efficiently derive meaningful insights from it. Cloud data lake storage like Amazon S3 has become the destination of choice for storing high volumes of data because it is inexpensive, scalable, and simple to manage. However, to analyze that data, companies have historically needed to move and copy that data into proprietary data warehouses — a process that is costly, complex, risky, and inflexible.

Dremio’s data lake engine sits between cloud data lake storage and data consumers, enabling them to directly query data for high-performing dashboards and interactive analytics without needing to copy the data into proprietary data warehouses, then subsequently needing to create aggregation tables, extracts, cubes, or other derivatives. Dremio also provides a shared semantic layer that empowers data analysts to discover, curate, analyze, and share datasets in a self-service manner, and centralizes data security and governance for data teams. The result is a simpler and more streamlined data architecture that reduces time to value while enhancing data security and eliminating vendor lock-in.

More broadly, open-source innovation and industry thought leadership is pivotal to the Dremio value proposition. For instance, Apache Arrow was originally our own internal memory format that we decided to open source. It’s now the standard for in-memory computing, with over 20 million downloads each month. More recently, we created Project Nessie, which brings Git-like version control to the data lake, accelerating the agility of data engineering, data science and analytics.

SaaS Factory: This week, you launched the Dremio Cloud. Tell us why you chose the unique architectural approach you took.

Tomer Shiran: Dremio Cloud is a cloud-native data lake query engine delivered as a service that scales with customer workloads. Increasingly, companies want fully-managed services that enable them to focus on deriving value from data instead of worrying about system setup and administration. So, developing a Dremio SaaS offering was a natural progression in our story.

Dremio Cloud provides high concurrency and low latency queries directly on Amazon S3, and a semantic layer that makes data consumable, consistent, and secure for analysts and data scientists. It consists of an always-on control plane that receives queries from clients and is responsible for query planning and engine management, and a data plane comprised of compute engines that are responsible for query execution.

The multi-tenant control plane is central to the Dremio Cloud customer experience, hosting all client-facing interactions, including the user interface, REST API, and data query endpoints. When a business user wants to run an analysis with Dremio Cloud, they connect their preferred BI tool such as Tableau, Power BI, SageMaker, Looker, or a Jupyter notebook, to the control plane at app.dremio.cloud. The control plane securely delegates query execution to compute engines automatically provisioned in the customer’s AWS account, so all data processing happens within the customer’s account.

The data plane architecture is comprised of multiple right-sized compute engines to support different workloads. Built on this multi-engine architecture, Dremio Cloud introduces the capability for engines to dynamically scale based on workload size, helping companies tackle any level of concurrency while maintaining consistent performance. All data is stored and processed within the customer account, and encrypted in transit and at rest, ensuring that customers have full control of their data. There are also no inbound connections into the data plane, so customers don’t have to poke holes in their firewalls/security groups. These features result in a stronger security and governance story for our customers.

Dremio Cloud architecture on AWS

The approach we took required significant technical innovation, including using end-to-end Apache Arrow to dramatically increase query performance. Without Arrow, serializing and deserializing data structures is inefficient, and results in wasted memory and CPU resources. Arrow allows Dremio to combine the benefits of columnar data structures with in-memory compute, providing performance benefits and the flexibility of complex data and dynamic schemas.

SaaS Factory: Who are your core customers, and how does this change with the introduction of Dremio Cloud?

Tomer Shiran: Dremio has always been designed to work for any company that wants to use its enterprise data strategically. Hundreds of enterprises across all industries use Dremio to power their cloud data lakes, including financial institutions like Standard Chartered Bank, pharmaceutical companies like Johnson & Johnson, and manufacturers like Honeywell. Amazon itself uses Dremio to analyze and power business intelligence on data in its internal data lake, such as supply chain data.

Even so, we have designed Dremio Cloud to be bi-directionally scalable, so that venture-backed startups that have a lot of data to be analyzed but don’t necessarily have the resources to operate their own data infrastructure or prefer not to burn hard-earned venture dollars on a cloud data warehouse, can use it effectively.

SaaS Factory: The addition of a SaaS product involves a comprehensive business and organization transformation. How have different functions of the organization evolved to better align with the SaaS business and delivery model?

Tomer Shiran: Indeed. On the product engineering front, we created hundreds of thousands of automated tests and a comprehensive CI/CD process. We expanded our product development organization to include site reliability engineering (SRE), DevOps and security teams, with leaders from companies such as Google and Salesforce. As a result, we’re now able to release updates to Dremio Cloud on a daily basis.

In addition to the changes on the product engineering team, we aligned our customer facing teams within the company to support a self-service adoption model. Our sales and marketing teams focus on driving high quality leads to the Dremio Cloud offering online, while our customer success and support teams leverage operational data and automation to provide proactive and targeted support to ensure strong customer satisfaction. The icing on the cake is that we can use Dremio internally on our own data as the foundation for this!

SaaS Factory: How did you engage AWS as you were developing Dremio Cloud?

Tomer Shiran: We’ve always had a special relationship with Amazon, partnering with multiple teams within the company. We work closely with numerous AWS services teams, such as Amazon S3, AWS Glue, and Amazon Lake Formation, to deliver integration between our services and to collaborate on new features. We partner with the AWS Marketplace team to distribute Dremio through the marketplace, enabling companies to consume our product while paying through AWS. Our sales and marketing teams work with the AWS Partner Network and AWS sales organizations to bring Dremio to AWS customers, thereby enabling AWS customers to build next-generation data lakes/lakehouses.

We’ve also had the privilege of working with the AWS SaaS Factory team over the last couple of years. When we embarked on our journey to build Dremio Cloud, we wanted to leverage state-of-the-art technology and best practices to create a best-in-class cloud service. As numerous SaaS and PaaS services have been built on AWS in the last 10 years, including AWS’ own services, we wanted to avoid the challenges and limitations that other services experience, while capitalizing on what worked well. To do so, we partnered with the SaaS Factory team to develop an architecture that delivers unparalleled scalability, security and performance and to develop a flexible usage-based pricing strategy to ensure an optimized SaaS delivery model for customers across all segments.

SaaS Factory: Dremio is now officially a ‘Unicorn’ based on your most recent series D funding round in January 2021. If you were speaking to aspiring founders, what would be your advice to them?

Tomer Shiran: If there’s a significant need in the market that you’re excited about, don’t hesitate to launch your own startup. But take the time to build the best product in that category. For instance, in the data infrastructure space, a significant amount of intellectual property is required to deliver a solid product. We spent a good five years at Dremio building what we thought would truly be a next-generation data lake engine from the ground up, with a focus on innovation and customer success. Once you build a strong foundation, exponential growth will be easier to realize. Today, six years after its founding, Dremio powers the cloud data lakes of many of the world’s largest enterprises and have raised over $200M in venture funding in the last year.

Dremio and AWS are excited about the future of data management, and the innovation we are delivering with Dremio Cloud. If you’re interested in learning more, please check out the Dremio Cloud page!

About AWS SaaS Factory

AWS SaaS Factory helps organizations at any stage of SaaS journey. Whether looking to build new products, migrate existing applications, or optimize SaaS solutions on AWS, the AWS SaaS Factory Program can help. SaaS builders and operators are encouraged to reach out to their account representative to inquire about engagement models and to work with the AWS SaaS Factory team.

Visit the AWS SaaS Factory Insights Hub to discover more technical and business content and best practices. Sign up to stay informed about the latest SaaS on AWS news, resources, and events.

Qbiq: Using AWS Lambda Container Images & Distributed ML to Optimize Construction

AWS Editorial Team — Fri, 16 Jul 2021 22:46:05 +0000

Guest post by Qbiq Team and Hilal Habashi. Startup Solutions Architect, AWS

Real estate software startup Qbiq system delivers an artificial intelligence (AI)-driven space planning design engine that generates large volumes of customized floor plans, compares alternatives, and optimizes the results. It instantly provides suggestions for the best layout design within the constraints of space utilization, costs, build time, efficiency, and a multitude of other factors. These non-trivial compute intensive calculations are done behind the scenes in a distributed manner by dividing each job request into multiple parts. More specifically, each plan received is broken down into sub plans, and each of these sub plans are processed using machine learning (ML) and nonlinear programming. First, we use ML models to recognize and analyze the different layout alternatives, and then we optimize the results using quadratic equations.

We chose AWS right at the start of our journey because of the flexibility and elasticity in AWS’s offering. It is a perfect fit considering our significant compute power requirements in each step of the pipeline. In fact, aside from our main workload, our preprocessing stage, which divides the plans and requires both spatial and geometrical analysis as well as a heavy hyper-parameters search – a non trivial workload by itself. We also heavily rely on its backbone network coverage and response times, its breadth of services, and the pay-as-you-go cost model. All of these allow us to scale our service and bring it to as many customers as we can.

In order to support our business model, we needed to provide our clients with a fast and reliable software-as-a-service (SaaS) solution, which is quite different from what we had in mind when we first started. We’ve had our fair share of challenges throughout the process of building the distributed workload presented in this post, and as with any case of significant architectural change, we needed to find a happy medium between the adoption of new technologies and onboarding and training our engineering team to build the best product we could.

Challenges

Developing distributed ML algorithms

We started out as a cloud-native SaaS solution running on Amazon EC2 instances. In order to perform a large number of computations in a limited run time (sub 5 minute), we had to divide our workload into tiny segments and run subtasks in a distributed manner. As we started to horizontally scale our compute on AWS, we also had to rework our algorithms, use a shared storage to store the different models we need to run, and build a queue of jobs received over time. Thanks to Amazon EFS, and Amazon SQS, we were able to quickly implement and maintain the supporting infrastructure for our distributed compute needs with little to no effort of our small team. EFS enabled us to share a storage between our AWS Lambda containers and use the drive as a shared mounted storage for our subprocesses.

Deploying and maintaining our algorithms

It was important for us to deploy our algorithms in a container-based approach without using proprietary tooling or having to write our own deployment scripts from scratch. We wanted a solution that would work with the traditional Git-ops approach and abstract out the environment that our containers ran on. We wanted to quickly scale out the compute power available based on demand and without worrying about provisioning or managing any container clusters. We chose to use Lambda containers with EFS over solutions based on AWS Fargate or EC2, with Amazon EBS since we wanted a highly flexible and scalable solution with a pay-as-you-go model that didn’t require us to perform container orchestration and management tasks.

Testing our algorithms at load

We had to write end-to-end tests in order to verify that our calculations performed correctly across our cluster of distributed tasks regardless of its size whether running it locally or on the cloud. In reality, when we ran and orchestrated our workload on EC2, we noticed that the more instances of distributed software we added, the harder it became to maintain the configurations and test our stack as a whole. Our high-level goal was to improve our cloud fidelity and avoid maintaining two sets of setup scripts with a mutual codebase. In addition, we wanted to be able to test our product at high load scenarios, without extensive configurations.

AWS Lambda Containers in our Cloud Architecture

The recently added support for custom container images in Lambda have ticked all the boxes for us. It was easy to write our algorithms and wrap them, along with all of our dependencies, inside a container image. After doing so, we had to mount volumes, add new model files onto our container images, and test our software both locally and on the cloud. Adopting containers has simplified our deployment process by solely relying on our trusted Git-based flow rather than on external devops tools. Finally, we managed to reduce our compute resource management to a simple task of configuring memory and vCPU for our Lambda, a task which was made even easier by Lambda Power Tuning. The container images running on Lambda in our environment are essentially the core of our product, and thanks to the mature state of Lambda event source integrations with the rest of the AWS services, the migration from EC2 was effortless.

High level overview:

We’ve built 3 Lambda functions using Amazon Elastic Container Registry (Amazon ECR) to store each of our 3 base images; one for each part of the process, with their basic dependencies and settings included as files and environment variables. We then connected EFS to store the outputs of each task so they will be rapidly available for the following tasks, while our distributed ML models were stored in Amazon S3. Since our distributed ML processes cut down the tasks even further, it was crucial to have a mutual file system that can be mounted and accessed within any container-based workload. This setup allowed us to share files with high throughput and low overhead between the various Lambda functions. We use several SQS queues containing information about the jobs, with incoming files.

This architecture allows us to achieve similar result and overall time to deployments over EC2 cluster while reducing costs by paying for only the time that the algorithms are running. For example, we are able to produce a batch of 50 results in 2 minutes by invoking 5,000 Lambda functions with maximum concurrent execution of 500, which results in an average cost of $1 per batch. On the other hand, running it on EC2 cluster with 500 parallel cores would have cost about twice as much and will necessitate provisioning the cluster. Finally, our ML engineers and data scientists can operate within the container-based environments they are familiar with while using one set of tests instead of writing a unique set for a non-containerized environment running on the cloud as they did before.

Managing our workload

In order to manage our distributed ML workload reliably, we implemented our own graph-based workflow manager. Each job request sent by a user starts out with a file being uploaded and its name being recorded as a Topic Name. We then create a sub-graph for each Topic Name that can contain multiple nodes under it – one node for each computational task to be processed. Each node in the job’s sub-graph contains an internal state representation. The edges in the subgraphs are a function of the input and the next task to be executed.

Our current workload is a 3-step process, meaning that we multiply the number of parts we divided each job into by three. The preprocessing pipeline breaks down the jobs and their corresponding files into multiple parts. A sub-graph is created for each job, and the workflow manager then loads the jobs into the system containing many nodes. We start processing the nodes inserted in parallel by traversing the edges of our graph and correspondingly loading the current state information into the next Lambda function, each step at a time. The workflow manager records the data in the corresponding node for each step of the process and part of the job and continues until all the data for the job is processed. Finally, we aggregate the results and communicate them to our customers.

The workflow manager communicates with the Lambda functions through a REST API using two types of POST requests:

1) `RequestRequired` is a synchronous request that can handle large payload size but has a response time limitation of 30 seconds.

2) `Event` is an asynchronous request and works well for sending multiple parallel jobs with a small payload.

Architecting the workflow manager:

This figure shows a schematic illustration of the workflow manager as a service provider to our backend system

The workflow manager sets up two helper threads: transmitter and consumer. The transmitter thread is responsible for checking the asynchronous jobs queue and managing which jobs to activate by sending them to the Lambda functions. We can control the number of simultaneous active jobs as a parameter in order to control active Lambdas concurrency. When a job is sent to be processed by a lambda function, the request contains a SQS queue name to which it should write the responses. The consumer thread is responsible for consuming the messages from the SQS queue and to distributing incoming job responses according to their topic.

As we deploy a large number of tasks with a limited worker pool so not to exceed concurrency limits, nodes that do not converge or require irregular computational time can significantly reduce the system performance. To protect against hanging or overly long jobs, a three timeouts system is used. The first timeout is the basic lambda timeout that is configured when setting up a Lambda function. This timeout should be pretty loose since it should potentially never be reached. The second timeout is within each lambda function. When the function handler is called, it starts an additional thread and waits for a specific amount of time. If the timeout is reached, the Lambda function sends back a timeout error response to the queue. The last is a topic timeout in the workflow manager that is agnostic to Lambda functions and is triggered if more than a certain amount of time is passed since the last job post was sent under this topic.

To achieve high performance with fast warmup (lower the start time), we used the Provisioned Concurrency feature of the Lambda functions. We also use an initialization call to warm up the Lambda function and prepare the data. The EFS throughput is provisioned to allow high MiB/sec to eliminate bottlenecks when multiple lambdas access the data stored on EFS in parallel.

Conclusion

At Qbiq, we bring cutting-edge AI, generative design, and optimization technology to real-estate planning. Using AWS Lambda image containers enables us to easily scale to hundreds of cloud processors loaded with hundreds of years of architectural experience, process the planning request, analyze different layout alternatives, and optimize the results.

We provide our customers with the best construction alternatives considering utilization, costs, build time, efficiency, and more. For more information about Qbiq, please visit our website, qbiq.ai.

Purple Ant on Using AWS to Transform the Insurance Industry

AWS Editorial Team — Fri, 02 Jul 2021 22:48:56 +0000

Guest post by Iman Nandi and Greeshma Nallapareddy, Solutions Architects, AWS

Purple Ant is a subscription-based property monitoring platform that enables its customers to detect, prevent, and track damage to their homes using IoT devices. Pankaj Parashar (CEO and co-founder of Purple Ant) founded Purple Ant after his insurance company recommended he install a car device that collects driving data to price his insurance. Parashar quickly realized a similar device for houses could not only help to accurately and fairly price home insurance, but also keep his family safe. This started his journey into leveraging IoT (Internet of Things) and AWS to transform the home insurance industry.

Pankaj Parashar, CEO & Co-founder, Purple Ant

How It Works

Purple Ant leverages IoT with AWS IoT Core, storage with Amazon S3 and Amazon DynamoDB, and data routing through Amazon API Gateway to monitor home health effectively and prevent damage based on data collected. Data is used to find signs of potential future damage and provide immediate solutions. Purple Ant lowers the cost of insurance by lowering the risk of the property getting damaged. This benefits both insurance companies who assume the risk and home owners by sending alerts on how to prevent damages, using sensor data to detect damages, and sending help to fix damages.

Leveraging AWS

Sensors from Purple Ant’s smart home devices collect data signifying the occurrence, or lack thereof, of a disastrous event like a water leak or fire in near real-time. As shown in the ingestion architecture diagram below, there is data flow established starting with collecting data from smart home devices connected to IoT Core. As data is collected, the IoT Rules Engine sends data to Amazon Kinesis Data Firehose for aggregating and organizing the data into various AWS data stores, including S3 and DynamoDB, using AWS Lambda functions.

As shown in the delivery architecture diagram, another Lambda function is utilized for sending alert messages and processed data to a React Native mobile application using API Gateway, as well as sending emails and text messages using Amazon SNS and Amazon SES. An Amazon RDS PostgreSQL database is also present for storing and delivering metadata. About 1.5 MB of data per home device is collected daily with an average of 10K expected devices to send data, totaling 15 GB of data inflow per day. In the future for scalability, the data warehouse solution Amazon Redshift will be utilized for data ingestion once the event volumes are large enough to impact the performance of the current setup.

Ingestion:

Delivery:

Enabling Scale and Differentiation

Manoj Narayanan, CTO at Purple Ant, says that leveraging AWS “dramatically increases incorporation of new features, allows global deployment, and scalable performance due to the ease of implementation available.” Using AWS IoT Core lets Purple Ant scale as they increase their user base while still keeping security top-of-mind. Lambda functions are triggered when sensor state changes are detected and allow them to trade capital expense for operational expense.

Insights

Throughout Purple Ant’s journey, the company has found that intentional choices need to be made on how to scale and cost optimize their business. Visualization is one critical aspect of their IoT functionality that needs to be responsive at scale. “We started with Superset and realized soon that it will lead into performance issues especially in an IoT event scenario. We ended up leveraging Redshift to ensure that the performance met expectations. So even within AWS, we have to choose the right technology option based on the specific application functional/non-functional need,” says Manoj. For scalability, a Redshift instance will be utilized to pull data onto their React dashboard rather than pulling data directly with Kinesis Data Firehose once event volumes are large enough to impact the performance of the current setup. When implementing in IoT Core, they found it easy to set up and use for their use case, eliminating the need to create any custom interfaces or integrations to make their architecture function as intended. Moreover, IoT Core was an out-of-the-box solution available for them to hook up to third-party hubs. “It was easy to create a ‘thing’ and establish a secure connection, and implementation went smoothly,” says Manoj. A flexible AWS service with standard implementation options enabled them to cover most of their scenarios.

To allocate costs where most necessary, Purple Ant maintains a large focus around their AWS spend. A large aspect of that is being able to choose the appropriate types of infrastructure to match their workloads. “Given that we are a startup, cost optimization is a high priority area for us. This means that we need to continuously keep track of the Amazon EC2 instances/sizes etc. amongst others to ensure that we are scaling to the right requirement,” says Manoj. The team was introduced by the AWS team to AWS Cost Explorer, which can be utilized for using reserved instances, taking advantage of the savings plans offered by AWS, and receiving other recommendations to reduce cost. “We use the Cost Explorer regularly to track the trend and adjust the costs where needed. The trend also helps us upgrade the infrastructure when needed so that performance doesn’t suffer,” says Manoj. Cost optimization is a continuous process for startups like Purple Ant that want to deliver the value of home monitoring to their customers in a successful way.

Conclusion

By leveraging AWS, Purple Ant is able to transform the way home insurance is priced, how home owners care for their properties, and how they can enable continuous monitoring to prevent home damage. Ultimately, homeowner premiums and repairs in the United States can be reduced significantly by up to 15-20% with their property monitoring devices, which can result in savings of approximately 16-22 billion dollars for consumers.

Syntegra: Bridging the Healthcare Data Gap

AWS Editorial Team — Wed, 30 Jun 2021 13:42:10 +0000

Guest post by Ofer Mendelevitch, co-founder and CTO, Syntegra

Data is at the heart of everything we do in healthcare: driving treatment decisions, research and development of new drugs, medical breakthroughs for rare disease patients and so much more. Yet, the inability to easily access and share high-quality, patient-level data inhibits us from leveraging its full potential to advance medical innovation and improve patient care.

These long-standing challenges led to the creation of Syntegra in 2019, after my co-founder Dr. Michael Lesh witnessed firsthand the roadblocks researchers face in accessing critical patient data to inform their efforts. I connected with Michael as he searched for solutions to these challenges, and we realized the wide array of use cases for synthetic data across the healthcare ecosystem. Pairing his industry knowledge and patient-focused perspective with my deep technical expertise was a no-brainer. Our mission at Syntegra is to leverage novel, generative deep learning models never before used in the healthcare space to bridge the healthcare data gap, enabling improved research and accelerating how treatments reach patients.

Current Challenges with Healthcare Data

Despite the emergence of big data in healthcare, it remains nearly impossible to share useful, comprehensive, patient-level data sets externally due to strict privacy concerns. Even internally, it can be cumbersome to share data with other teams for analytic and educational purposes.

Amid the expanding focus in healthcare on Real World Data (RWD) for innovation and throughout the drug development lifecycle, real-world data sources require ever-increasing levels of privacy guarantees. However, the current methods of de-identification to protect patient privacy are largely inadequate for two reasons; first, they are vulnerable to re-identification attacks, resulting in continued privacy concerns; second, the de-identification process results in a significant reduction in data quality due to obfuscation of data, addition of noise and removal of small cohorts.

The Syntegra Solution

This clear market opportunity led to Syntegra becoming the first company to use the power of transformer-based language models like GPT-2 and GPT-3 in healthcare to create comprehensive, synthetic datasets that maintain full statistical fidelity and patient privacy. The result is a dramatic increase in availability and usage of healthcare data, enabling tremendous value creation across the healthcare continuum.

Our unique synthetic data engine trains on the underlying structured healthcare data and can then generate new synthetic patient records that contain no link to the original patients yet maintain all statistical properties of the original dataset. This approach allows Syntegra to generate patient-level healthcare data that can be used as a privacy-preserving replacement for real data in both simple and complex analytics, such as building predictive models or conducting survival analysis. In addition, this approach enables the generation of enhanced data that goes beyond the limitations of the real data. For example, the Syntegra engine does not remove small cohorts, which are often so important for research in rare disease. Furthermore, it enables bias normalization and imputation of missing values, ensuring the resulting synthetic dataset is fit for purpose and a significant leap in quality for real world data.

Syntegra’s approach allows health systems, life sciences, payers and commercial data providers to seamlessly share privacy-guaranteed healthcare information internally and externally without the need for expensive and time-consuming compliance and contractual structures, secure sandboxes and complicated access protocols. Privacy is guaranteed in a way that goes beyond HIPAA and GDPR compliance. And with statistical fidelity preserved at the patient-level – unheard of in healthcare – Syntegra-generated synthetic data can be used immediately for statistical analysis, reporting and predictive models. But to build the necessary trust that Syntegra is accomplishing these goals, measurement and proof is needed. Syntegra has developed a set of metrics to assess both statistical fidelity and privacy of synthetic data to help demonstrate the performance of its synthetic data engine.

Leveraging AWS to Empower Growth

As an early-stage startup, we leverage a number of AWS technologies at Syntegra to help us be a better partner to our growing and diverse customer base.

Harnessing the computing power of Amazon EC2 allows us to effectively and efficiently scale our synthetic data engine with customer workloads. We use the AWS python SDK, boto3, to dynamically create GPU-enabled P3 and P4 EC2 instances based on our custom Amazon Machine Image/Launch template. Our AMI has pre-configured access permissions, the latest cuda drivers and fabric managers (for the new A100 GPU instances), and a pre-loaded version of our docker container with all of the necessary installed python packages installed to minimize the startup time for all jobs. Amazon’s streamlined authentication process for accessing EC2 makes it easier for us to work across different machines and in different settings, letting us work nimbly for a wide range of healthcare stakeholders and across a variety of use cases.

For basic data storage, we use the cost-effective Amazon S3 to preserve all outputs from our EC2 machines even after the machine has been terminated. The Amazon Elastic Container Registry allows us to continue to update our container images without affecting the rest of our workflow, so that we can work across a range of datasets and organizations.

Powering our synthetic data engine through AWS allows us to work smarter and more efficiently while also being cost-effective as Syntegra continues to grow and scale. The offering of spot instances lets us keep our costs low, while the AWS API allows us to easily query for the real-time price of spot and non-spot instances. Syntegra has leveraged EC2 spot instances to reduce costs by up to 80%, when compared to the previous use of on-demand instances. Additionally, during the training of a large language model, Syntegra detects when a spot instance will be terminated by AWS and starts a new spot instance to resume exactly where the previous instance left off. This allows us to save the cost of cold restarts, leading to saving many hours of instance uptime.

Looking to the Future

AWS’ wide range of offerings not only enables us to leverage the services that fit our company’s needs right now, but it allows us to look to the future and think big. Our goal of democratizing healthcare data through the use of synthetic data promises to revolutionize the way we think about accessing and sharing healthcare data, and subsequently the potential to change the lives of countless patients.

Curious to learn more about synthetic data generation in healthcare? Visit syntegra.io to learn more and connect with us at contact@syntegra.io.

Changing the Landscape of Molecular Testing

AWS Editorial Team — Mon, 28 Jun 2021 15:42:36 +0000

Guest post by Paul Flook, PhD, CIO and VP, Software Engineering, ChromaCode

This past year has demonstrated, now more than ever, the critical need to be able to develop and deploy rapid molecular testing at scale. The ability to do this has emerged as a major differentiator for ChromaCode, a startup diagnostics company based in Carlsbad, California. We focus on developing Real-Time Polymerase Chain Reaction (RT-PCR) tests that can detect multiple genes or pathogens in biological samples.

Before the onset of COVID-19, our focus had been on tackling other important challenges such as tick-borne pathogens and drug resistant bacterial genes using our proprietary High-Definition PCR (HDPCR) platform1. Our unique advanced multiplex technology is well-suited for addressing the ongoing research needs due to rise of tick-borne diseases or the 2.8 million cases of antibiotic-resistant infections in the US per year. But testing for these targets has taken a back seat as most nations have been consumed with COVID-19 testing.

In the spring of 2020, just as the scale of the pandemic was becoming clear, ChromaCode worked with a lab at University of California San Diego to prototype a COVID-19 molecular test2. As this progressed and we saw the problems faced by labs ramping up their development efforts, we realized there was a significant opportunity for us to help with COVID-19 testing. A critical enabler in the resulting effort was the design of our software, ChromaCode Cloud, which made it possible to quickly deliver the data interpretation and reporting components of the test. A working assay was ready within about a week from the start of the project, and we are proud to have since shipped more than four million COVID-19 tests in less than eight months, despite the numerous roadblocks that inherently plague diagnostic testing.

Rapid roll out of new molecular tests is a key challenge

Labs don’t have the time to validate, install, and maintain new software and hardware, and most organizations strive to keep their investment in these to a minimum. Together these can represent significant barriers to adoption of new diagnostics in labs seeking to offer the best tests at the lowest cost.

In addition to the technology itself, there tend to be various challenges in delivering a solution that can be easily integrated into a lab’s infrastructure. There are compliance and data privacy issues with diagnostic products. Maintaining compatibility with critical operating system updates may require updates that customers need to install and validate themselves. But research labs cannot spend all their time certifying software updates. The list goes on.

A unique cloud-based solution for medical devices

ChromaCode’s HDPCR platform changes the conversation around molecular testing. The technology supports a simple workflow that enables users to upload standard PCR instrument output files to our software platform using a web application. Data is automatically analyzed and results reported to the end user. Providing a 4x expansion of the analytical capacity of most PCR instruments, labs can leverage their existing PCR instruments, maintain their existing protocols, and can scale at a higher rate.

Our technology uses digital signal processing, enhanced PCR reagents, and cloud-based analytics. It’s this unique combination that enables us to significantly increase test capabilities while decreasing costs, factors that contributed to being named as a Main Track Winning Team in the XPRIZE Rapid COVID Testing competition. Over 100 entrants participated in this competition which was designed to recognize the best low cost, high throughput COVID tests. Our success emphasized how ChromaCode’s patented technology is changing the PCR landscape, and how ChromaCode Cloud enables increased scale.

ChromaCode Cloud is comprised of a web application, analysis tools, and multiple platform services running in a Kubernetes cluster and using AWS foundational products such as Amazon EC2, Amazon RDS, and Amazon S3 to deliver an end-to-end analysis solution. Our use of AWS from the start avoided the need for us to build and operate our own core infrastructure and allowed us to rapidly scale as the COVID testing demand exploded. We have also built a separate informatics platform that enables us to analyze aggregate de-identified analysis and operations data to gain insights into the performance of our tests.

Because of this, with nominal cost deployment, we can better support our customers with the cloud, which leverages the agility, security, breadth and depth of AWS services to enable real-time access to the data required to support a wide range of accurate and affordable molecular diagnostics, including COVID-19 testing.

The high availability of the system means we can deploy new maintenance and security updates immediately without disruption to customers; there’s no need to stop testing or revalidate security updates, for example. With effectively unlimited storage capacity, we can free customers from concerns about data archiving. In terms of compliance and security standards, AWS has industry-leading capabilities, superior to anything a typical customer would implement themselves.

By leveraging a cloud-based software solution, not only are we able to provide an improved experience, but labs are able to easily adopt and rapidly scale. And, we now have an incredibly rich data set. We are currently analyzing aggregated de-identified data that allows us to look at the performance of our assays, gain better insights into our inventory, and even offers us the potential to look at epidemiological trends.

Expanding our testing capabilities to address new challenges

The ongoing challenges around COVID-19 mean that an important next step for ChromaCode is an expansion of our testing capabilities to support new demands on labs at the frontline. A primary objective has been to develop a new test that can identify additional respiratory viruses such as the seasonal flu and Respiratory Syncytial Virus (RSV), whose symptoms may resemble COVID-19. We are currently planning the extensive clinical trials necessary for this new test to be cleared by the FDA as a full In Vitro Diagnostic test.3

We are also developing research use only tests to identify new COVD-19 variants3. The emergence of new strains that have different etiologies and transmissibility pose significant risks to controlling the pandemic. Genomic sequencing has been used effectively to characterize new strains, but it is impractical to apply this at population scale. Instead, rapid development of PCR-based screening assays and tests for variants may be critical to understand large scale epidemiological changes.

Not only is ChromaCode Cloud imperative to support these tests, but our ability to aggregate de-identified test data in our informatics platform also means that we are able to undertake large scale monitoring of different variants that can provide unique insights.

If you wish to learn more about ChromaCode, our products or our software you can visit chromacode.com, or request a copy of our white paper from info@chromacode.com.

1 For Research Use Only. Not for use in diagnostic procedures.

2 For Emergency Use Authorization Only. For in vitro diagnostic use. For prescription use only.

3 The content provided herin may relate to products or workflows that are still in development, have not been officially released or fully validated and is subject to change without further notice.

How Alloverse is Creating a More Inclusive Internet with AWS

AWS Editorial Team — Tue, 22 Jun 2021 15:21:20 +0000

Guest post by Nadja Lindgren, Nordic Scale Account Manager, AWS & Sophia Granfors, Startup Marketing Manager Nordics, AWS

The adoption of augmented and virtual reality (AR/VR) technologies have risen in recent years within marketing, education, healthcare, tourism, events, fashion, entertainment, retail, gaming, and other sectors. However, due to high development costs, the complexities of AR/VR devices, and other challenges, these technologies are still in their early phases of adoption.

Co-Founded by CEO Julie Despraz, Swedish startup Alloverse has developed an open-source platform for virtual collaboration that is being used to build the spatial internet. The company’s platform and tools enable users to create virtual workspaces and 3D applications to populate them. We recently sat down with Despraz to chat VR, sustainability, and celebrating even the smallest of wins.

Describe Alloverse’s vision and its product offering.

We are developing a collaborative open source platform with a marketplace for productivity apps in VR. We provide tools and infrastructure that enable users to write VR apps that run server-side using a networked API, and join virtual worlds on the same terms as other users. In other words, they can compose environments in which they and their colleagues can work on interoperating apps at the same time.

We want to shape the next 50 years of the internet in a more inclusive manner than it would have evolved otherwise. So, everything we do is open source, distributed, and federated. We work in the open (an open and public Discord, constantly sharing progress on social media). We’re writing open specifications that we hope will become the foundation of a new era of internet protocols, replacing or augmenting the 2D web. Furthermore, we believe that the future of software has to be collaborative by design so that we can always be and work together, especially as we move into spatial software.

How will virtual and augmented reality affect sustainability?

VR has had a major, cross-industry impact in areas such as training, upskilling teams, and keeping people mentally sane and physically fit. COVID-19 showed that although we’re far from a mainstream VR/AR world, many businesses have benefited from it, particularly when it comes to training employees. Perkins Coie surveyed nearly 200 professionals last year spanning startups, enterprise technology firms, and investors for insights on the trajectory of the immersive technology industry. Some 70% of respondents believe that organizations will focus on adopting VR/AR in training and development during 2021. In the healthcare sector, 68% of the respondents believe that AR/VR training simulations will be the primary focus of new solutions and applications through 2022.

What has been the biggest challenge building your startup thus far?

We’re building an open-source infrastructure/platform for VR/AR apps from scratch, which is difficult, fundamental work. Instead of picking an existing platform such as Unreal or Unity, we decided to build our own using Lövr and other languages, such as Lua. We’re spending a lot of time learning about how to build VR infrastructure that will be safe, robust, and easy for developers to use.

What are three pieces of advice you’d give to early-stage startups?

Build a startup with people you trust and care for. Make sure that you share the same values and dreams for your venture. If things don’t go to plan, remain friends. Enjoy the ride, have fun, and celebrate every win, no matter how small! When forced to make an important decision such as a new hire, or signing a deal or a terms sheet, turn to peers or who have gone through the same experience. There’s nothing better than real-world advice. And if you need to use lawyers, use them for one day (eight hours) in sync with lawyers from the other side. This will cut costs and make the process for both parties more efficient.

How has AWS supported your startup journey?

The AWS Startup Team proactively checks in with us and puts energy into making sure we have everything we need, whether that’s through architectural support or joint marketing. Their interest in our business is much appreciated.

Do you want to get started and accelerate your business with AWS? Visit aws-startuploft-emea.com for exclusive events and support.

Nadja works on AWS Nordic Startup Team helping early-stage startups scale and grow with AWS. Prior to her time at AWS Nadja led the brand transformation of Hamilton, a Swedish business law firm, establishing content marketing, social media strategies and digital rebranding.

Sophia is a Stockholm-based startup marketer passionate about enabling early-stage startups to succeed. Prior to AWS, she was helping entrepreneurs with financing through angel investor matchmaking. She has a strong passion for diversity and inclusion, to level the playing field for underrepresented founders.