Bio

I enjoy turning AI research into production systems. My work lives at the intersection of stochastic optimization, AI, and machine learning applied to high-stakes physical world problems. Throughout my career I’ve carried ideas the full distance, from a theorem on a whiteboard to production software making decisions on billions of dollars worth of freight per year. I work across the stack from doing reesarch to prove why an approach works, building the pipeline that ships it, and leading the team that takes it to market.

Starting with the research, I hold a PhD in Electrical Engineering from Princeton University, where I specialized in stochastic optimization under Professor Warren B. Powell at the CASTLE Lab (Computational Stochastic Optimization and Learning Lab). My doctoral work centered on multi-agent sequential decision-making for complex physical systems under uncertainty. We designed collaborative strategies for agents with complementary skillsets to pursue shared objectives in dynamic, stochastic environments. It gave me a durable way of thinking about decisions under uncertainty that I’ve applied to commercial problems ever since.

After my PhD, I joined Optimal Dynamics, a startup founded by Warren and Daniel Powell, to commercialize CASTLE Lab’s network optimization and reinforcement learning algorithms. As Director of Artificial Intelligence, I led the Statistics and Machine Learning group through five years of growth from early stages to a Series C (over $100M raised) backed by top tier investors like Koch Disruptive Technologies and Bessemer Venture Partners.

I work across the full technical and product stack to connect research to production engineering. I led the development of multiple flagship AI products at Optimal Dynamics which serve dozens of enterprise customers and manage thousands of trucks per day.

I am currently on a short research sabbatical, doing independent research on post-training and reinforcement learning for large language models while I weigh my next step. I’m drawn to hard problems where rigorous methods meet real-world stakes.