- Temporal Representation Alignment: Successor Features Enable Emergent Compositionality in Robot Instruction Following (2025)
- Horizon Generalization in Reinforcement Learning (2025)
- Accelerating Goal-Conditioned RL Algorithms and Research (2025)
- Learning to Assist Humans without Inferring Rewards (2024)
- Inference via Interpolation: Contrastive Representations Provably Enable Planning and Inference (2024)
Research Experience
Currently conducting research at Berkeley Artificial Intelligence Research (BAIR), supported by an NDSEG fellowship.
Education
Ph.D.: University of California, Berkeley, advised by Anca Dragan and Sergey Levine; B.S.: Stanford University, in Computer Science and Mathematics, advised by Dorsa Sadigh.
Background
Research interests include reinforcement learning, human-AI interaction, and robotics. Before coming to Berkeley, he got his bachelor's degree in Computer Science and Mathematics at Stanford University, where he worked with Dorsa Sadigh.
Miscellany
Contact information: Email, Google Scholar, LinkedIn, GitHub, Twitter, Bluesky