🧬💻 Issue #129: Harnessing mRNA's Full Potential & Preparing APIs For The Agentic Era
Sagar Batchu shares his thoughts on the impact of AI agents for software engineering and how developers can start preparing. Michael Swift & Kerna Labs launch to build better genetic medicines.
Good morning & happy Thursday. First and foremost, our thoughts and prayers are with our friends and families in LA who have been impacted by these unprecedented fires. It has been a trying time in Los Angeles over the past week, and we hope those reading are safe.
This week, we explore two sectors where AI is being heavily disruptive: APIs & biotech therapeutics. StoryHouse portfolio founder Sagar Batchu shares a guide on preparing APIs for the AI agent era, and biotech startup Kerna Labs and Claremont graduate Michael Swift launch from stealth with a $6M Seed round and an investment from StoryHouse. It’s a Claremont world out there. 👇
~ Josh, Miles, Matthew, Pat
📢 👥 Community Voices: It’s Time To Start Preparing APIs for the AI Agent Era
Sagar Batchu (HMC’15) started his career as a software engineer for IoT proptech startup Enlighted, where fellow Claremont alum Joe Costello was the CEO. After Siemens acquired Enlighted, Sagar left to join LiveRamp’s new London office and grew into a Director of Engineering role, helping build and scale the company’s London engineering team. Now, Sagar is the CEO and co-founder of StoryHouse portfolio company Speakeasy — a developer-first API DevEx platform that enables developers to offer best-in-class self-service experiences for API consumers. They recently secured a $15M Series A round led by Wesley Chan and Pegah Ebrahimi at FPV Ventures, with continued support from previous investors GV and Quiet Capital.
With agentic AI, integrations are no longer static and immutable concepts.
The software engineering world has always designed APIs with human developers in mind, who have predictable and consistent usage patterns. But a distinct new kind of API consumer is rapidly emerging: the AI agent.
Unlike human developers, AI agents behave like a thread. They are task-based bots spun up to achieve a specific goal. Completing their task might require them to form API connections on-demand, query an API, and close the connection. With agentic AI, integrations are no longer static and immutable concepts. Distinct, short-lived integrations will become the norm. And with AI agents becoming the norm within dozens of coding assistant tools and platforms, the question is no longer ‘if’ but ‘when’ API design approaches will change along with them.
How AI Agents Will Change API and Application Development
Limitless API Calls
Let’s use a flight booking application as an opportunity to examine how AI agents are already changing application development:
Here, the difference in API usage boils down to deterministic vs. non-deterministic. The traditional app gathers input from the user and makes a single API call to surface the results. Usually, a single action by the user corresponds to a single API call.
Compare that to an AI agent application, which theoretically has no limit on the number of calls it can make from a single user request. It will continue to iterate and make several or even dozens of calls until it has satisfied the user’s request. If the result is unsatisfactory, the user can also re-prompt the agent.
Granular Security for AI Agents
On the security side, connection patterns and authentication flows will transform. In the short term, AI-based apps should increase the number of instances where an application requests access to a 3p API on behalf of a user. In the long term, the dynamic nature of AI agents might produce more sophisticated approaches to security.
For example, imagine an AI agent that needs to access your calendar to book a flight. Rather than requesting blanket calendar edit access, it might receive a temporary “micro-scope,” allowing it to check availability for dates relevant to travel planning. These micro-scopes would expire once the task is complete, minimizing the security surface area. As a model for LLMs, OAuth will likely die in a year.
Managing Context Across APIs
Traditional APIs treat each request as isolated, but AI agents consider the bigger picture when maintaining conversation context across many calls.
As AI agents increasingly need to coordinate between multiple services to complete tasks, the real challenge is managing context across different APIs. Where will that context live? And who needs to know about it? Currently, that context lives with the LLM, which picks and chooses the pieces it needs to share when sending an API request. But it isn’t hard to imagine a future with standardized “context blocks,” such that context can be shared and accessed across several API providers….
Check out Sagar’s full article to learn more about how developers can prepare for the new challenges and opportunities of the AI agent era.👇
🚨Claremonster Call-Out: Michael Swift & Kerna Labs
Julia Peng, Melissa Moore, Amit Deshwar, and Claremont alum Michael Swift (CMC ‘16) are the co-founders of Kerna Labs, which is building foundation models of RNA to unlock the full potential of mRNA as a therapeutic modality beyond vaccines. The AI-powered approach allows Kerna Labs to optimize mRNA like never before. Their platform solves critical bottlenecks in mRNA payload design and delivery, including potency and protein output, tissue specificity, half-life, and manufacturability. The team has already demonstrated mRNA sequences that perform magnitudes better on these parameters.
Kerna Labs is built and backed by the leading minds in machine learning and mRNA. Julia brings eight years of experience in deep tech partnerships, go-to-market strategies, and B2B sales from her work in biomaterials. Melissa, a former Moderna executive, played a key role in the development of the first mRNA medicines. Amit was the founding ML scientist at Deep Genomics, where he led a team of 45 building RNA models. Michael is a Claremont graduate and a Stanford PhD grad in systems biology. He is an expert in high-throughput biology and single-cell analysis, with over a decade of RNA biology experience, including as a Research Fellow at Longitude Capital.
Kerna Labs recently emerged from stealth with a $6M seed round backed by Susa Ventures, Gradient Ventures (Google’s AI-focused fund), and StoryHouse. We are excited to be an investor in Kerna Labs and their team of top biotech veterans in AI and mRNA. 👇
💼 Who’s Hiring?: Kerna Labs, Terra AI, & Fastbreak.ai
StoryHouse portfolio company Kerna Labs, co-founded by Claremont grad and founding scientist Michael Swift (CMC ‘16), builds foundation models of RNA to unlock the full potential of mRNA as a therapeutic modality beyond vaccines. Their platform solves critical bottlenecks in mRNA payload design and delivery, including potency and protein output, tissue specificity, half-life, and manufacturability. Kerna Labs recently emerged from stealth with a $6M seed round backed by Susa Ventures, Gradient Ventures, and StoryHouse. They are seeking an ML/Computational Biology Scientist to join their growing team. Reach out to us if you are interested!
StoryHouse portfolio company Terra AI is building an AI platform for critical minerals and clean energy development. They're leveraging AI modeling, reasoning, and optimization to enhance the performance, predictability, and safety of reservoir, wind, and mineral development projects. Claremont gradAnthony Corso (HMC '14) andJohn Mern(who previously worked for KoBold Metals, founded by CMC grad Kurt House)are the co-founders of Terra, and they are looking to hire an experienced ML engineer to lead the development of critical parts of their platform. If you’re interested in applying your engineering skills to solve some of the most critical problems in energy, check out the openings below or reach out to us:
StoryHouse portfolio company Fastbreak.ai is a rapidly growing leader in sports technology. The company combines artificial intelligence with sports expertise to transform how leagues, teams, and brands operate and collaborate. Some of the most prestigious sports leagues and industry giants use their cutting-edge SaaS scheduling platform, including the NBA, NHL, MLB, the SEC, Adidas, and more. Claremont alum Timothy Carnes (HMC ‘05), John Stewart, and Chris Groer are the co-founders of the company, and they are hiring for several open roles:
Check out the other ~2,000 open jobs at 290+ Claremont-affiliated companies here on our Storyboard. Plus, create a profile and enter your preferences to get alerted to new job postings relevant to you, be they the 350+ remote jobs, 150+ internships, or 500+ part-time positions available. We’ve published research that shows that Claremont-founded companies that disproportionately hire Claremont talent outperform — so pay attention, Claremonsters!
If any of these roles catch your eye 👀 , apply and mention StoryHouse Review. Or, if you are an employer looking to hire tip-top Claremont talent, fill out this form to have your jobs featured.
🗣️ Conversations on the Interwebz:
This week’s must-watch 📺
In his recent interview, Claremont alum Jerry Chu (CGU ‘18) discussed how his startup, Lofty AI, is revolutionizing how we think about owning and investing in real estate. He shared the inspiring story of Lofty’s evolution from a SaaS platform to a real estate exchange and explained how tenants can now become co-owners of their rental properties. Jerry is the founder and CEO of Lofty AI, a marketplace that lets you invest in U.S. rental properties starting at just $50. Lofty has raised over $6M in VC funding from investors like Y Combinator, Hustle Fund, Rebel Fund, StoryHouse, and more.
This week’s Claremont financing 💸
Claremont grad Kurt House (CMC ‘01) is the co-founder and CEO of unicorn KoBold Metals, a mining company that uses artificial intelligence to mine rare earth elements. After securing $491M last October, KoBold recently announced that it has raised a total of $537M in Series C funding at a $2.96B valuation. With this financing round, KoBold has raised a total of ~$1B, which will be used to advance its exploration technology and position itself as a key player in the mining sector.
Congratulations to Claremont entrepreneur Alexei Agratchev (CMC ‘99) and RetailNext for securing a majority growth investment from Battery Ventures. Alexei is the co-founder and CEO of RetailNext, a global leader in smart store analytics for optimizing store experiences. The recent investment will support the expansion of RetailNext’s international team, accelerate product innovation, and provide significant capital for acquisitions. In his recent LinkedIn post, Alexei also shares his key learnings and thoughts on this significant milestone for his company. 👇
💡 The Answer To AI’s Struggles In Software Development…..Claremont grad Jared Cosulich (CMC ‘02) recently shared his thoughts on how AI is changing software development and its challenges with understanding complex systems due to limited context. Jared also introduces the concepts of isolation and abstraction as tools to make AI more effective in navigating and contributing to software projects by breaking applications into manageable parts and summarizing essential details. Jared is the CTO and co-founder of software simulator startup CodeYam.
⚽ Fastbreak.ai Partners With The NWSL…..Sports technology company Fastbreak.airecently announced a strategic partnership with the National Women’s Soccer League, the premier women’s professional soccer league in the world. The NWSL selected Fastbreak’s platform for its scheduling processes, with the goal of enhancing operational efficiency and delivering an improved experience for teams and fans alike. In addition to the NWSL, Fastbreak is also working with other industry giants, such as the NFL, NBA, Adidas, and Apple. Claremont alum Timothy Carnes (HMC ‘05) is the co-founder and CTO.
