Letian Wang 王乐天

PhD student at University of Toronto, supervised by Steven Waslander. We're interested in perception and decision-making algorithms that can generalize and adapt to the open world.

Student Researcher at Google DeepMind, with Misha Andriluka, Andrei Zanfir, Eduard Bazavan, João Carreira, Andrew Zisserman, Kaiming He. I am working on unified text-video models.

Research intern at MiroMind, with Jifeng Dai and Lidong Bing. I worked on LLM deep research agents, long-term memory, long context and efficient sparse attention. Contributor to MiroThinker, the state-of-the-art open-source deep research model.

Research intern at NVIDIA Research, with Peter Karkus, Seung Wook Kim, Boris Ivanovic, Yue Wang, Sanja Fidler, and Marco Pavone. I worked on self-supervised 3D-aware visual pretraining via generalizable neural radiance field [NeurIPS'24].

Research assistant at the Robotics Institute of Carnegie Mellon University, with Changliu Liu and Yeping Hu. I worked on generalizable motion prediction algorithms in different scenarios, and social interaction for autonomous driving [NeurIPS'21, AAAI'22, Book]

Research assistant at UC Berkeley, with Masayoshi Tomizuka, Liting Sun, and Wei Zhan. I worked on socially-compatible behavior generation for autonomous driving [RA-L'21] (Best Paper Award Honorable Mention)

Research assistant at the HKUST RI, with Fu Zhang. I worked on unified control strategy for vertically take-off and landing UAVs.