James Bowman
Chief Technology Officer AI Proteins
James W. Bowman, Ph.D., is a member of the founding team and Chief Technology Officer at AI Proteins, where he helped build the framework powering the company’s discovery engine. He previously trained as a Postdoctoral Fellow at the Institute for Protein Innovation, Boston Children’s Hospital, and Harvard Medical School, developing novel miniproteins for cancer therapeutics that led to multiple patents and licenses. With more than a decade of protein engineering experience, James has contributed to projects ranging from GPCR stabilization for nanobody discovery during his Ph.D. to the creation of a miniprotein tumor antigen binder now in clinical trials. While at AI Proteins, he has been named on patents covering immune cell engagers for oncology, inhibitors for autoimmunity, and agonists for metabolic disease. James earned his Ph.D. in Genetics, Molecular, and Cellular Biology from the Keck School of Medicine of USC and his B.A. in Neuroscience from the University of Southern California.
Seminars
This workshop will feature deep dive sessions on:
From Design to Optimization: Building Integrated Platforms for Functionalizing De Novo Proteins
- Addressing the unique challenges of producing and testing designed proteins
- Transitioning from designing for functional exploration to designing for optimization
- Coordinating multidisciplinary teams to measure diverse properties and accelerate design-build-test-learn cycles
Building Curated Databases to Fine-Tune Property-Prediction Models for Designed Proteins
- Bridging the wet- and dry-lab divide to prioritize and plan data generation
- Assessing workflow maturity and data quality before model fine-tuning
- Designing curated datasets that support reliable prediction of protein properties
- Where has computational design actually moved the needle, and how are you proving it internally to speed up drug pipelines?
- How to control for data quality and biological reality in ML-driven design loops? And what specific changes have you made to your data generation strategy to make it truly usable for training models?
- How are you deciding what to prioritize today across de novo design, optimization workflows, and engineering, given they sit at different points in the hype cycle?
