Machine Learning Scientist, Biomolecular Structure Prediction

We're looking for a motivated and creative Machine Learning (ML) Scientist to drive research into biomolecular structure prediction. You will build and improve all-atom co-folding models, with a focus on the complexes that matter most to our programs: antibody–antigen interfaces and protein–nucleic acid assemblies. These remain among the hardest cases for current models, and closing that gap directly accelerates our antibody design and genome editing work.

The ideal candidate has trained structure prediction models at a leading academic or industry lab and knows where today's models break. You will be a self-directed researcher who moves quickly from hypothesis to trained model to rigorous evaluation. As an early member of the ML team, you will help shape how structure modeling fits alongside our generative models (ProGen3, E1) and experimental platforms.

• →Develop and train all-atom structure prediction models, improving accuracy on antibody–antigen and protein–nucleic acid complexes
• →Diagnose failure modes of current models (e.g., CDR-H3 and epitope placement, interface geometry, nucleic acid conformations, confidence calibration) and design architecture, training, and data interventions to address them
• →Curate large-scale training data from structural databases, distillation sets, and Profluent's own experimental binding and function data
• →Integrate structure models with our generative models for design and filtering, and work with protein design teams on experimental validation
• →Present results to colleagues in regular update meetings

• PhD (or equivalent industry experience) in computer science, computational biology, biophysics, or a related field
• Hands-on experience building and training biomolecular structure prediction models (e.g., AlphaFold- or Protenix-style co-folding models) at a leading academic or industry lab
• Publications at major ML conferences (NeurIPS, ICML, ICLR) or top science journals, or widely used open-source models
• Experience with deep learning frameworks such as PyTorch or JAX, including multi-GPU training

• Experience modeling antibody structure and antibody–antigen docking, including CDR loop modeling and epitope prediction
• Experience modeling protein–DNA or protein–RNA complexes
• Deep familiarity with structural data pipelines: mmCIF processing, MSA and template search, structural clustering, and dataset splitting
• Experience with diffusion or flow-matching models for structure generation or structure-conditioned design
• Experience with large-scale distributed training and cloud compute
• Familiarity with experimental structural and binding data (cryo-EM, crystallography, display-based binding assays)

Legal authorization to work in the United States is required. In compliance with federal law, all persons hired must verify their identity and work eligibility and complete the required employment verification form upon hire.