My research interest lies in multitask/continual learning, reinforcement learning, imitation learning and their applications in robotics. Specifically, I am interested in designing theoretically sound algorithms and neural architectures for training large-scale multi-purpose agents.