I am a principal researcher in the Reinforcement Learning group at MSR Redmond. My general research interests lie in the theory of decision-making under uncertainty and its applications, ranging from building AI for autonomous sailplane UAVs to designing algorithms for Bing’s next-generation Web crawler. Most recently, I have been working on representation learning for RL with a special focus on robotic manipulation.
I received a CS Ph.D. from the University of Washington – Seattle and a double B.A. in CS and applied mathematics from UC Berkeley.