Translate decades of protein knowledge into designs that bend chemistry to our will.

The role

We are looking for a self-directed leader looking to push the frontier of ML for enzyme design, focusing on data, representations, and architectures that capture sequence-structure-dynamics-function relationships.

• Develop novel approaches to modeling our proprietary simulation-derived data.
• Develop conditional generators that propose designs satisfying catalytic geometry, stability, and manufacturability constraints.
• Design optimization loops that explore and improve the design space efficiently.
• Ship production-grade code: models that run on our cluster today, not “interesting ideas” parked on a branch.
• Collaborate with simulation physicists and wet-lab scientists to close the design-build-test-learn cycle in weeks, not quarters.

You might be a fit if you

• PhD (or equivalent experience) in Computer Science, Computational Biology, Biophysics, or a related discipline.
• Strong publication or project record in top ML or life-science venues (NeurIPS, ICML, ICLR, Nature, Science, Nature Methods, etc.)
• Proven track record applying deep learning to protein sequence–structure–dynamics-function problems (e.g. property prediction, structure prediction, de novo design, directed evolution, etc).
• Strong software-engineering skills and fluency in PyTorch or JAX.
• Familiarity with modern generative approaches (diffusion and flow matching models).
• Understand practical wet-lab constraints: expression systems, thermostability, assay variability.

Bonus points

• Experience with equivariant models, graph neural networks, and other 3D modeling approaches.
• Contributions to open-source protein ML libraries.
• Hands-on background in ML-driven protein engineering.
• An understanding of molecular biology, biocatalysis and/or protein science.