I'm an ML researcher working on training and verification systems for open-ended scientific discovery. I lead Dynamical Systems, infrastructure for agentic materials labs. We're building scientific intelligence for materials workflows. Planning, search, verification, and judgment become trainable loops around human scientists, models, and physical labs. The longer goal is to change how humans and AI explore the edges of scientific discovery together.
Previously, I co-founded Arc Intelligence, where we built ATLAS, a continual-learning system for production agents. ATLAS turned deployed trajectories into reward-scored memory, inference-time guidance, and distillation data, and became the foundation for my current work on behavior-to-signal-to-training loops for scientific agents. Before that, I built RL environments and distributed training infrastructure at NEAR, and helped grow open-source developer ecosystems at Protocol Labs.
I've come to believe that the best coaches are great teachers, and the best teachers are great learners. I built a career in machine learning by asking questions. That started when I was a football coach at Ohio State, Clemson, and the Los Angeles Rams, where I learned how to teach athletes to perform under pressure and how to construct environments where feedback, difficulty, and trust produce growth. This led me to a different arena as an early-stage investor at Emerson Collective, an Assistant Professor at NYU, and a doctoral student at the University of Illinois Urbana-Champaign, where I studied learning design and how to keep learners at the edge of learnability.
The science of learning is often counter-intuitive. Inverting discovery to be open-ended means shifting from goal-oriented, hypothesis-driven methodologies to approaches that prioritize curiosity, exploration, and the continuous generation of new problems, which has been the throughline of my approach to doing meaningful work and building a better world for my daughter to grow up in.