Lila is building a platform where AI and automation co-evolve to solve the hardest problems in medicine. Within Life Science AI (LSAI), the Foundation Models team builds foundation models that learn across biological sequence, molecular structure, and experimental data to power automated scientific discovery across Lila's life science domains.
We are seeking a Scientist I or II to work on structure prediction and co-folding. The team's current emphasis is protein–protein and complex prediction in support of antibody and biologics design, and on making those predictions good enough to drive real experimental decisions. You will contribute across problem formulation, model design, training, evaluation, and integration into Lila's closed-loop discovery engine.
This is an IC role for someone building deep expertise in structure-aware generative AI for biology. You will own research sub-problems end to end, collaborate closely with experimental scientists to close the computational–experimental loop, and contribute to Lila's presence in the broader scientific community.
What You'll Be Building
Train and evaluate structure prediction and co-folding models for protein complexes, protein–protein interactions, and related biomolecular systems
Build and extend models informed by AlphaFold-style co-folding, diffusion models, protein language models, and related structure-aware ML methods
Build rigorous evaluation frameworks to ensure model generalization to challenging de novo design problems
Scale training, inference, and evaluation workflows across large GPU clusters
Be part of the end-to-end ML process within Lila's “Lab-in-the-Loop” lifecycle: shape data generation strategy, build pipeline models, and design feedback loops where experimental results improve model performance
Contribute to adjacent foundation model research where it strengthens the structural work, including biological sequence design and multimodal scientific reasoning
Translate biological questions into well-defined ML problems and interpret model outputs alongside wet-lab scientists, structural biologists, and computational biologists
Support research quality and methodology standards within the foundation models program
Qualification
PhD in Computer ScienceProficiency in ML frameworks (PyTorchAntibodyContributions to open-source ML tools
Required
PhD in Computer Science, Machine Learning, Computational Biology, Biophysics, or a related quantitative field (or Master's with equivalent research experience)
Hands-on experience training deep learning models on molecular, protein, or structural data
Strong foundation in generative model architectures and training, with demonstrated ability to design careful experiments, ablations, and evaluations
Ability to formulate and execute research independently, from problem definition through experimentation
Familiarity with at least one life science domain (structural biology, protein engineering, molecular biology, genomics, or related)
Experience collaborating with experimental scientists or working with biological/chemical data
Proficiency in ML frameworks (PyTorch, JAX, or TensorFlow) and experience with GPU-based training workflows
Preferred
Experience training or extending co-folding, structure prediction, protein–protein, or diffusion deep learning models
Experience with AlphaFold or AlphaFold-derived methods (e.g., Boltz, Protenix), RFdiffusion, or protein language models
Antibody, biologics, or protein design experience, including structure-guided optimization
Familiarity with distributed training infrastructure and large-scale scientific data pipelines
Contributions to open-source ML tools, frameworks, or benchmark datasets for scientific applications
Experience with active learning loops or closed-loop experimental workflows
Experience integrating ML models into agentic scientific workflows
High-impact publications or open‐source contributions in AI for Science in relevant venues (NeurIPS, ICML, ICLR, AAAI, Nature Methods, Nature Biotechnology, or equivalent)
