Projects

Efficient One-Step Flow Matching for VLA Models

Nov 2025 - Present | ScaleLab, Shanghai Jiao Tong University

Developing efficient one-step flow matching visuomotor policies for Vision-Language-Action (VLA) models. This research aims to improve the efficiency and performance of robotic manipulation tasks through advanced machine learning techniques.

Supervisor: Professor Yao Mu

UniRoboSim: Universal Robot Simulation Framework

Apr 2025 - Oct 2025 | GIFT Future Scholar Plan, SJTU

Developed a universal, high-fidelity, task-oriented robot simulation framework for reinforcement learning research. The framework supports multiple robot platforms and provides a unified interface for RL algorithms.

Supervisor: Professor Qin Tong

Status: Paper under review