David Baker

Nobel-laureate protein designer who built the computational toolkit (Rosetta, RoseTTAFold, RFdiffusion, ProteinMPNN) that anchors modern generative biology.

1. Identity & Background

David Baker is a professor of biochemistry at the University of Washington and director of the university's Institute for Protein Design (IPD). He shared the 2024 Nobel Prize in Chemistry — with Demis Hassabis and John Jumper (DeepMind) — for computational protein design and prediction; Baker's half recognized de novo protein design, Hassabis/Jumper's half the alphafold3 structure-prediction lineage [1].

His lab is the origin point for a large fraction of today's generative protein-design tooling: Rosetta (physics-based design), RoseTTAFold (structure prediction), RFdiffusion (generative backbone design via SE(3)-equivariant diffusion), and ProteinMPNN (inverse folding — designing a sequence for a target backbone). The RFdiffusion → ProteinMPNN pipeline is the field's canonical "geometry then sequence" workflow (see aifs-biology) [local].

2. Currently Doing

Baker continues to lead IPD and the broader push from static structure prediction toward generative design — inventing proteins with new functions (binders, enzymes, vaccines, materials) rather than only predicting natural ones. His lab's open-science posture (releasing weights and methods) has made its tools the shared substrate for much of the AI-for-biology startup ecosystem [local].

3. Companies & Projects

Institute for Protein Design (UW) — director; the lab whose spinouts and open tools seeded the AI-drug-discovery wave.
Xaira Therapeutics — Baker is a co-founder of this $1B-seeded AI drug-discovery venture (launched 2024, Arch Venture Partners + Foresite Labs); Xaira's X-Cell virtual-cell work also sits in the AI Virtual Cell landscape tracked in ai-virtual-cell.
Chai Discovery — the open-model antibody/structure startup (Chai-1, Chai-2) was founded by researchers from Baker's UW lab; see chai-discovery.

4. Unique Observations

One lab, two Nobel halves. Baker's design work and DeepMind's prediction work are two sides of the same protein problem; the 2024 prize split institutionalizes the prediction/design duality that now structures the whole ai-for-science biology field.
The academic-industrial bridge. The week's research focused specifically on which of the ~10 leading AI drug-discovery companies have direct ties to Baker's lab — the answer being several via founder lineage (Chai Discovery) and co-founding (Xaira), confirming that IPD functions as an industry-wide talent/methods source, not just an academic group [local: 2026-08-24-deepopencode.md].
Open weights as moat-in-reverse. Baker's lab releasing RFdiffusion/ProteinMPNN/RoseTTAFold freely accelerated competitors, but also made his methods the default baseline every startup must beat — a recurring theme in aifs-biology's "general vs specialized" debate.

Sources

[1] Nobel Prize in Chemistry 2024 — David Baker facts (nobelprize.org)
[local] 2026-08-24-summary.md — daily research summary
[local] 2026-08-24-deepopencode.md — DeepOpenCode session log (David Baker + company-collaboration queries)