Director, AI/ML - Generative Foundation Models for Biotherapeutics

At Lilly, the work is demanding because patients are waiting. We unite caring with discovery to help make life better for people around the world, knowing that every decision, every detail, and every day matters. Headquartered in Indianapolis, Indiana, our over 50,000 employees around the globe take on complex challenges to discover and deliver life-changing medicines, strengthen how health is understood and managed, and support the communities we serve. This is hard, urgent, selfless work—but it’s work worth doing. If you’re driven by purpose and ready to bring your best to work that truly matters for patients, we invite you to join us. Director, AI/ML - Generative Foundation Models for Biotherapeutics At Lilly, we unite caring with discovery to make life better for people around the world. For more than 25 years, Lilly's Biotechnology Discovery Research (BioTDR) organization has advanced novel antibody and peptide therapeutics from concept through clinical development to market in areas of high unmet medical need. This role is based at the Lilly Biotechnology Center in San Diego, where we are expanding our computational capabilities to build a generative design platform. By bringing together AI/ML, protein engineering, automation, structural biology, and high-throughput experimentation, BioTDR is creating an integrated de novo design platform capable of generating, ranking, and optimizing novel therapeutic antibodies and other biotherapeutics with improved activity, safety, manufacturability, and target coverage. Designs are rapidly evaluated through experimental validation, creating a continuous learning cycle that strengthens both models and discovery outcomes. The program is supported by dedicated capacity on LillyPod, our wholly owned 1,016-GPU NVIDIA Blackwell Ultra SuperPOD. Help us push the de novo protein design frontier beyond binders to molecules that can become medicines, faster. Ready to make an impact? Join us. Primary Responsibilities Our Generative Design platform has three pillars: Generative Foundation Models , Generative Protein Design & Optimization , and Active Learning & Design Validation . The Director of Generative Foundation Models will lead the development of protein and antibody foundation models and validate their performance against experimental outcomes, not just published baselines. You will be accountable for the program's generative model families end to end, across architecture choice, training strategy, and validation, and co-define the program's technical agenda with leadership. You will lead a team of AI researchers and ML infrastructure engineers while staying hands-on and giving direct technical guidance on every model in your workstream. This is an opportunity for a scientific leader who wants to build and scale technologies that directly shape therapeutic molecule discovery. · Lead generative AI and protein foundation model development , driving state-of-the-art machine learning approaches for representation learning, generative protein design, sequence-structure-function modeling, and multimodal biological learning. · Set training strategy , developing the pre-training and fine-tuning approach for your models across architecture choice, model family, training curriculum, and evaluation. · Devise training infrastructure , designing and steering the training lifecycle and scaling approach against LillyPod capacity, accountable for reproducibility, checkpointing, and throughput, with the ML infrastructure engineers on your team. · Direct confidence and calibration work , treating it as a first-class research objective and reducing the gap between model-ranked and experimentally validated design performance. · Drive model improvement , resolving underperforming models to root cause across architecture, data, and execution, and making timely decisions to retrain, pivot, or discontinue. · Drive the advancement of foundation model portfolio , establishing success metrics and ensuring progress translates into measurable experimental outcomes. · Leverage multimodal and proprietary biological data , creating differentiated learning advantages by bringing proprietary sequence, structure, binding, functional, developability, and other biological measurements into model development. · Lead a high-performing team and provide scientific and technical leadership , recruiting and developing AI researchers and engineers on your team and providing sustained technical mentorship. Guide model architecture choices, experimental design, technical prioritization, and research direction while remaining sufficiently close to the science and technology to challenge assumptions and identify new opportunities. · Evaluate and incorporate external innovation , staying current with rapidly evolving advances in foundation models, generative AI, protein design, structural modeling, and related technologies. Evaluate external methods and collaborations and determine when to build, adapt, or partner. · Communicate scientific strategy and impact , presenting technical progress, experimental validation, key learnings, and strategic recommendations to scientific leadership and broader R&D stakeholders. Contribute to publications, external collaborations, and scientific presentations where appropriate. Basic Requirements · Ph.D. in computer science, mathematics, physics, computational biology, or a related quantitative field, with 3&#43; years of relevant research experience following the Ph.D. (or an M.S. with 6&#43; years), including experience as technical lead of a multi-person modeling effort. · First-author or equivalently attributable contribution to structure-based generative model. Plus, a track record of carrying projects end to end. · Experience training deep learning models on multi- node infrastructure; strong Python and PyTorch. Additional Preferences · Depth in equivariant architectures, diffusion or flow matching on structure, and graph transformers. · Experience with inverse folding, side-chain packing, all-atom generation, or conformational ensembles. · Work on uncertainty quantification or confidence prediction validated against experimental outcomes, not only against held-out structures. · Contribution to a de novo binder or antibody design effort that reached experimentally validated designs. · Antibody or VHH experience is desired but not required. · Experience improving model efficiency, building smaller or faster models at equal accuracy. · Experience mentoring junior AI researchers or ML engineers. <span st