Robot Learning · Manipulation
Ph.D. Student in Electrical and Computer Engineering at the National University of Singapore
I work on robot learning and robotic manipulation, with an emphasis on reinforcement learning, vision-language-action models, simulation, and real-robot systems.
I received my M.Sc. in Robotics from the National University of Singapore and my B.Eng. in Robotics Engineering from the University of Electronic Science and Technology of China.
Goal: Build reliable robots that can learn, adapt, and improve through real-world interaction.
Conference on Robot Learning (CoRL), 2026
A lightweight visual comparator that turns one successful demonstration into a dense reward while suppressing uncertain progress estimates that may cause reward hacking.
Conference on Robot Learning (CoRL), 2026
A closed-loop real-to-sim-to-real framework that reconstructs cloth state from RGB observations and refines nominal actions through FLASH rollouts and prior-guided MPPI.
arXiv preprint, 2026
An agentic framework that turns a single in-the-wild image into a simulation-ready deformable asset through a generate–simulate–diagnose–refine loop, using simulated material responses to repair upstream generation errors.
arXiv preprint, 2026
A vision-language agent framework that converts a real-world recording of object–robot interaction into a simulatable episodic twin, supporting downstream policy fine-tuning and evaluation.
Built VLA data-collection and real-robot rollout infrastructure, reproduced real-world RL baselines, integrated human intervention, and evaluated dense progress rewards on dual-arm manipulation tasks.
Led the electrical subsystem of the national-champion elephant and bunny robots, owning the control architecture, distributed real-time communication, state estimation, and field-oriented actuation stack end to end.