Harvard in Tech AI Edition: Ankit Gupta, Visting Partner at Y Combinator and Head of AI at Ginkgo Bioworks
Ankit Gupta is a Visiting Partner at Y Combinator and the Head of AI at Ginkgo Bioworks, a biotechnology company offering tools and services for cell programming across industries such as agriculture, pharmaceuticals, and chemicals. He joined Ginkgo in February 2024 following the acquisition of Reverie Labs, a small molecule therapeutics company that leverages machine learning to develop oncology treatments. Ankit co-founded Reverie Labs with his college blockmate, Jonah Kallenbach. Both were AI enthusiasts with experience in biology and a shared vision of creating computational models to improve drug development. Reverie was founded shortly after their college graduation and became part of Y Combinator in early 2018.
Reverie initially aimed to build software for pharmaceutical companies but quickly encountered challenges in proving its value. Traditional metrics like R-squared scores failed to resonate with potential clients. Recognizing the need to demonstrate their technology’s capabilities, the team pivoted to becoming a fully integrated pharmaceutical company. By running their own drug discovery pipeline, they showcased the results of their models firsthand. To bridge the gap between biology and computation, they assembled a team that included computer scientists, small molecule experts, and chemists. This approach allowed Reverie to partner successfully with industry giants like Roche and Genentech, securing milestone-based agreements. Their strategy focused on presenting the intrinsic value of the molecules they developed, enabling potential partners to evaluate the assets independently of the company’s reputation.
In 2024, Ginkgo Bioworks acquired Reverie Labs. Familiar with Ginkgo’s history as the first biotech company to participate in Y Combinator, the Reverie team was drawn to its commitment to enabling other biotech companies to grow. The acquisition allowed Ginkgo to accelerate its AI efforts without building everything from scratch. Ankit transitioned to leading Ginkgo’s AI team, which recently launched its first Model API. This tool gives customers access to Ginkgo’s AI models through a user-friendly portal, marking a significant step in democratizing AI for biotech. Additionally, Ginkgo unveiled Datapoints, a platform designed to address the critical need for large-scale, model-ready datasets in biotechnology. More information about Datapoints is available here.
Reflecting on his time at Reverie, Ankit describes the experience as humbling and filled with rapid learning. Coming from a computer science background with limited formal training in biology, he quickly developed expertise in the field. He highlights the importance of interdisciplinary learning, noting that STEM fields often share analytical frameworks that make it possible to transition between domains. Whether computer scientists learning biology or biologists exploring computation, success requires exposure, practice, and a willingness to embrace what is unknown.
On the transition from startup founder to corporate leader, Ankit observes that building a company is one of the fastest ways to learn. Startups evolve rapidly, often requiring founders to adapt to a new company every six months. As a founder, he explains, every decision feels existential, with employees’ livelihoods and investors’ capital depending on the company’s success.
Reflecting on a decade in machine learning, Ankit notes that from 2013 to 2020, AI primarily served as an enabling technology, improving processes within established business models. For instance, Google enhanced restaurant recommendations with AI, and Facebook improved photo tagging. However, with the advent of breakthroughs like GPT-3, AI has become transformative, challenging core business models and reshaping how software is monetized and distributed. While there is uncertainty about which business models will ultimately succeed, tools that simplify AI integration, like Ginkgo’s Model API, are expected to play a pivotal role. Ankit also finds agent flows to be an area of growing interest.
In the biotech sector, Ankit emphasizes the need for more data. Existing datasets are insufficient given the complexity of biology and are often designed for hypothesis testing rather than training machine learning models. Effective model training requires millions of data points, far more than traditional experiments provide. Unlocking AI’s potential in biotech will depend on generating large-scale datasets, a challenge Ginkgo aims to address with Datapoints.
For students and aspiring founders, Ankit advises focusing on building strong relationships. College provides a unique opportunity to connect with brilliant peers, and these connections often have lasting value. He also encourages students to gain technical proficiency, particularly in STEM fields like computer science, as these skills are foundational for solving complex problems and driving innovation. Much of what he learned in computer science classes at Harvard continues to shape his work at Ginkgo.
In January 2025, Ankit is starting a role as a Visiting Partner at Y-Combinator, the startup accelerator where Reverie had its roots. At YC, Ankit will work with hundreds of startup founders across the full spectrum of technological sectors, and he’s excited to work with the next generation of founders to build new companies. He emphasizes that it is a great time to build a startup — with the advancement of AI tools, a nimble team can do more than ever before, and Ankit is excited to advise companies as they expand in this era.
