Tarun is a third year PhD student at University of Oxford, currently interning with autonomous vehicle research group at NVIDIA. His thesis focuses on improving the sample efficiency of cooperative multi-agent reinforcement learning (MARL) methods via transfer learning, role-based learning and solving MARL pathologies like relative overgeneralisation.

During his internship, he will be working on leveraging large semantic models for novelty in RL. For more information, please see his personal website.