Research
I’m a postdoctoral researcher at CLEAR Lab. Currently, I focus on rigorous robot learning system design:
- Egovideo-based controller design for dexterous hand;
- Reinforcement learning post-training for embodied foundation models;
- World (action) models.
I am actively recruiting RAs! Please send emails to zengx3@sustech.edu.cn!
My PhD research focuses on statistical fundamental limits and the relevant algorithm designs of learning for dynamics and control, including Koopman observable learning [To Appear], co-stabilization [CDC 2026], direct data-driven control [ACC 2025, TAC 2026, O. Hugo Schuck Best Paper Award], system identification [L-CSS 2025 and CDC 2025], and learning to stabilize [CDC 2023]. You can check my PhD proposal slides for a detailed overview.
Feel free to send me emails!