# Surge Biswas

> Surge Biswas is the co-founder and CEO of Nabla Bio, a Cambridge, Massachusetts biotech that designs antibodies with AI. As a graduate student in George Church's Harvard lab he helped pioneer protein language modeling, a technique now central to machine-learning-guided protein design. At Nabla he built JAM, a generative system he calls 'molecular auto-complete' that designs therapeutic antibodies from a target sequence alone, including against notoriously undruggable membrane proteins like GPCRs. The company has raised roughly $37M and struck collaborations with Takeda, AstraZeneca, and Bristol Myers Squibb worth over a billion dollars in potential payments.

- **Role:** Co-founder & CEO at Nabla Bio
- **Organizations:** Nabla Bio, Harvard University, Wyss Institute
- **Nationality:** United States
- **Education:** B.S., University of North Carolina at Chapel Hill, Master of Philosophy (MPhil), University of Cambridge, PhD (George Church Lab), Harvard University
- **Known for:** Helped pioneer protein language modeling during his Harvard PhD, a foundational technique for modern protein design, Co-founded Nabla Bio and built JAM (Joint Atomic Modeling), a generative AI system for de novo antibody design, Demonstrated the first de novo antibody designs relevant to therapeutic discovery against soluble and multipass membrane proteins, including GPCRs

## Career timeline

- **2015** — Begins PhD at Harvard University in George Church's lab
- **2019** — Co-authors early protein language modeling work (Nature Methods)
- **2020** — Earns PhD; co-founds Nabla Bio with Frances Anastassacos; joins Y Combinator S20 batch
- **2021** — Raises $11M seed round from Khosla Ventures and Zetta Venture Partners
- **2022** — Begins collaboration with Takeda on next-generation biologics
- **2024-05** — Closes $26M Series A led by Radical Ventures; announces AstraZeneca, Bristol Myers Squibb, and Takeda collaborations
- **2024-11** — Reports AI-designed antibodies binding the GPCR CXCR7 and de novo designs against membrane proteins
- **2025-10** — Signs second Takeda collaboration, a deal worth potentially over $1 billion in success-based payments

## Achievements

- Helped pioneer protein language modeling during his Harvard PhD, a foundational technique for modern protein design
- Co-founded Nabla Bio and built JAM (Joint Atomic Modeling), a generative AI system for de novo antibody design
- Demonstrated the first de novo antibody designs relevant to therapeutic discovery against soluble and multipass membrane proteins, including GPCRs
- Showed test-time scaling ('introspection') improving SARS-CoV-2 spike binder success rates 22-fold over a single design pass
- Raised approximately $37M in total funding and secured collaborations with Takeda, AstraZeneca, and Bristol Myers Squibb
- Published protein language model research in Nature Methods and Nature Biotechnology

## Latest updates

- **2025-10** — Nabla Bio signed a second collaboration with Takeda to advance AI-driven design of protein therapeutics, with potential success-based payments exceeding $1 billion.
- **2025-01** — Posted a bioRxiv preprint on de novo design of epitope-specific antibodies against soluble and multipass membrane proteins.

## Links

- Website: https://nabla.bio
- LinkedIn: http://www.linkedin.com/in/surge-biswas-a8b61270
- Twitter/X: https://twitter.com/nablabio
- YouTube: https://www.youtube.com/watch?v=7_YDkVPI464

---

Profile page: https://yespress.io/surge-biswas
Published by YesPress — https://yespress.io
Last updated: 2026-06-03