My main research interest is deep reinforcement learning, although recently I’ve also done work in sim-to-real robot learning, and have interests in imitation learning and model-based RL. I like thinking about what’s necessary to make deep RL applicable to more real-world problems.