Overview
Dayhoff Labs reverse engineers the chemistry of life using AI and experiments. In this role, you will build and apply foundational models of enzyme catalysis with use cases in pharma manufacture, agriculture and industry. As a member of the AI team you will build, maintain and use frontier AI models on real world problems together with our partners in industry and academia.
Role
Computational Scientist, Biocatalysis. Dayhoff Labs · London or Manchester
At Dayhoff Labs we reverse engineer the chemistry of life using AI and experiment. In this role, you will build and apply foundational models of enzyme catalysis with use cases in pharma manufacture, agriculture and industry. As a member of the AI team you will build, maintain and use frontier AI models on real world problems together with our partners in industry and academia.
The role
The role is pragmatic computational enzyme engineering. You will both build and use internal models, and work with and fine tune external models like Boltz, RFdiffusion and LigandMPNN. You’ll also be happy to use classical biocatalysis methods where appropriate. The role is to augment both discovery and optimization of enzymes, working closely with our wet lab engineers and foundation model team to deliver solutions for the biocatalysis community.
What you’ll actually be doing
- Owning the computational side of one to three commercial projects at a time, end to end: substrate analysis, starting-point selection, optimisation strategy, design rounds, in silico characterisation.
- Picking the tool stack per project. There is no fixed pipeline. You defend your choices on technical grounds and revise them when the data says otherwise.
- Designing and training novel architectures
- Acquiring new data for training both computationally and experimentally.
- Working closely with the wet-lab team on assay design, hit-call thresholds, and iteration.
What we’re looking for
Role-description signals matter more than CV signals. We care about:
- A track record of putting computational designs into wet labs and tracking what happened. The ratio of worked-to-didn’t-worked is less important than whether you can explain the failures mechanistically.
- Fluency across the protein ML stack (At least some of ESM, AlphaFold or Boltz, RFdiffusion, ProteinMPNN / LigandMPNN, docking) and a similar fluency in biocatalysis fundamentals (mechanism, kinetics, common cofactors, expression bottlenecks).
- Good taste in tool selection.
- Comfort talking to chemists and fermentation engineers about your model choices in their language.
- A sceptical disposition toward your own outputs. You will tell us when a prediction is junk and you will not over-claim.
Practicalities
Very competitive compensation. The role could be based in London or Manchester UK. We sponsor visas where the case is strong.
To apply
Send a CV and a short note — no cover-letter format required — describing the most recent computational design you put into a wet lab, what happened, and what you would do differently. Email careers@dayhofflabs.com with subject line “Computational Scientist, Biocatalysis.”
#J-18808-Ljbffr…