Postdoctoral Associate at Duke University · Department of Surgery
I develop reinforcement learning and control methods for decision-making in healthcare, and the oversight that keeps adaptive, multi-agent AI systems within safe limits.
Previously, I was a Research Scientist at the University of Chicago, building patient-specific deep RL for adaptive proton therapy. I hold a PhD in Computer Science from Emory University, where I was advised by Lars Ruthotto and supported by a Google PhD Fellowship.
- Oversight of clinical AI: supervisory control for adaptive and multi-agent systems
- Offline decision-making: safety-constrained learning from clinical data
- RL for treatment planning: patient-specific deep RL for radiation therapy
- Closed-loop biomedical control: neuromodulation and glycemic control
- Optimal control + ML: neural ODEs, Pontryagin's Maximum Principle, HJB equations
Website · LinkedIn · Email · Google Scholar