Research
My research is organized around three connected directions: AI agents in the economy, organizations, and science; foundation models and learning for operational systems; and experiments, markets, and games.
Foundation Models and Learning for Operational Systems
I build reusable models that learn the dynamics of operational systems such as queues, arrivals, inventories, and service processes, simulate them, and transfer decision strategies across related problems. This agenda connects generative simulation, reinforcement learning, queueing systems, and off-policy evaluation. Earlier simulation and reinforcement-learning papers are methodological foundations for this agenda, not papers retroactively relabeled as foundation-model research.