Research spans reliable ML and decision-making: uncertainty quantification and calibration, online/stochastic optimization and game‑theoretic learning, causal and decision‑focused methods, and theory for LLMs and test‑time scaling.