AI-native pharma company, bringing new treatments to patients faster and more efficiently.
New York, NY
Joined November 2014
- Tweets231
- Following30
- Followers2.4K
- Likes107
Pinned Tweet
The world has more promising discovered drug candidates than ever. The problem is developing them. We built Formation Bio to fix that, using AI to develop drugs faster and with a higher likelihood of success.
Thank you @amyfeldman and @Forbes telling the story of our mission, and our vision for the future of pharma.
forbes.com/sites/amyfeldman/…
(1/3) Phyla is Formation Bio’s context layer for causal biology. We benchmarked it against Claude Science and GPT-Rosalind on writing full research dossiers for 7 drug target–disease pairs, and the results are in. 99.7% of Phyla’s statements carried a citation, compared to Claude Science: 87% (Claude Science) and 86% (GPT-Rosalind). Phyla never dropped below 99.3% on any pair.
(2/3) In an audit of claims, Phyla led on every target-disease pair. 92% of Phyla’s claims were fully backed by the source it cited, vs. 80% (Claude Science) and 67% (GPT-Rosalind).
(3/3) Phyla aggregates 90 tools across 27 biomedical sources and ties every single claim to a checkable citation. When asked a plain language question, Phyla pulls the gene record, checks clinical precedent, computes tissue-specific expression, and runs mechanistic analysis over the results — returning an answer with every claim traced to a source. Here’s how we built it, and how it’s used at Formation: formation.bio/blog/phyla
Phyla aggregates 90 tools across 27 biomedical sources and ties every single claim to a checkable citation. Benchmarked against two other frontier AI life-science platforms, writing full research dossiers across 7 target–disease pairs: 99.7% of Phyla’s statements carried a citation (vs. 87% and 86%), and 92% of its claims were fully backed by the cited source (vs. 80% and 67%). See how we built it, and how it’s being used at Formation: formation.bio/blog/phyla
“As a company in the era of AI, you should have a thesis for what predictions will happen first, what will happen last, and make sure your company strategy absorbs this as a net tailwind versus a headwind.”
Our thesis: AI will improve drug discovery first, toxicity next, and clinical efficacy last. Our CEO @BenjamineYLiu explores this in his Uneven Frontiers essay – link in comments.
Formation Bio made #7 on @FastCompany's Best Workplaces for Innovators 2026.
This piece highlights one of our internal systems: ARK (AI-ready Repository of Knowledge), our in-house AI platform that connects all internal systems and data across the company. Full write-up here: fastcompany.com/91591991/for…
Our open roles: formation.bio/careers
Learn more about ARK: formation.bio/blog/ark-mcp-g…
Proud to be selected as #7 on @FastCompany's 2026 100 Best Workplaces for Innovators, plus as a winner on the Biotech, Healthcare & Life Sciences list.
Find us on the full list: fastcompany.com/best-workpla…
We're hiring across New York and Boston now: formation.bio/careers
Our CEO @BenjamineYLiu recently sat down with @LanceGlinn on @NYSE's Inside the ICE House to talk about how and why Formation Bio is tackling the drug development bottleneck, a problem that is only set to grow as the number of drugs at the top of the funnel increases. Links to listen in comments:
Hear from @formationbio co-founder & CEO Ben Liu on @ICEHousePodcast where he talks about...
- Bottleneck's slowing down drug development
- AI's impact on drug discovery
- Bringing new medicines to patients
Spotify: open.spotify.com/episode/60l…
YouTube: youtu.be/DaN84yEwBvU?si=tOHz…
Formation Bio retweeted
"In a world of drug discovery abundance, the scarce capability is no longer making drugs, but rather, knowing which drugs are worth developing."
@BenjamineYLiu, co-founder and CEO of Formation Bio, explains how AI will transform the economics of the pharma industry - link in replies
What drew Mike Ehlers, our new Chief Scientific Officer and Head of R&D, to Formation Bio? Hear it in his own words.
More on Mike’s life outside the lab in today’s @statnews executive moves column: statnews.com/pharmalot/2026/…
In his latest blog, our Principal Engineer Kenny Peng revisits the Unix philosophy (do one thing well, compose freely, shared interfaces) as a blueprint for agentic system design.
Using biopharma as the test case, he breaks down why hardcoded pipelines break down over time, and what composable alternatives require. Read here: linkedin.com/posts/kpeng_the…
Formation Bio retweeted
.@BenjamineYLiu has raised more than $600M for his AI-native biopharma company - and he's hiring!
@formationbio is applying AI to drug development, with a mission to bring new treatments to patients faster and more efficiently.
The team is hiring across the board, including:
- Director, Brand & Communications
- Senior Director or VP, Product
- General Manager (Drug Asset Lead)
- Principal Data Scientist, Medical Imaging
- Data Scientist, Portfolio Optimization
- Director, Business Development - Search & Evaluation
Full story + list of open roles in thread!
Formation Bio retweeted
Formation Bio said it has appointed former Biogen research head Michael Ehlers as its chief scientific officer, bringing a seasoned drug and biotech veteran into the AI-focused health startup. bloomberg.com/news/articles/…
Formation Bio retweeted
Excited to share that Mike Ehlers is joining @formationbio as Chief Scientific Officer and Head of R&D. Over the course of his career as a scientist, drug developer, and investor, Mike has advanced 40+ drugs into the clinic, with approvals spanning Alzheimer's, ALS, MS, hemophilia, and SMA. Welcome to the team, Mike!
The @decodingbio AI x Bio Summit lands in NYC this week, and our CEO and Co-Founder @BenjamineYLiu will be speaking at the @awscloud Executive Reception that follows, joining a panel on what it takes to scale AI in pharma.
Read more and request to join: luma.com/iylrgtia
This Thursday, our Co-Founder & CEO @BenYLiu takes the stage at @decodingbio's AI x Bio Summit at the @NYSE to discuss how platform companies are creating new business models in biotech – alongside voices from @Nabla Bio, @Manifold Bio, and @Dimension Capital. Link to register: events.decodingbio.com/summi…
