Building robots that perceive, reason, and act reliably in the physical world.
I build robotics systems that connect learning-based intelligence with real-world deployment. My work spans embodied AI, robot learning, VLA/VLM models, reinforcement learning, perception, planning, control, embedded systems, and the ROS2 engineering stack.
Currently, I work on edge AI and robotics applications at D-Robotics, focusing on deploying LLM/VLA workflows on RDK platforms. I am also an incoming MEng student in Electrical and Computer Engineering at the University of Waterloo for Fall 2026.
| Research Direction | Engineering Stack | Current Focus |
|---|---|---|
| Embodied AI, VLA/VLM, robot learning, world models | ROS2, SLAM, planning, control, STM32, ESP32, RDK X3/X5/S100 | Edge deployment of LLM/VLA workflows for real robots |
| Project | Focus | Highlights |
|---|---|---|
| gemma4-e2b-rdk-s100p | Edge LLM / RDK S100P | Deploying large-model workflows on D-Robotics edge hardware. |
| meta_sim2024_ep_docker | Sim2Real / Robotics Docker | Environment support for robotics simulation and deployment workflows. |
| BDX Bipedal Robot Replication | RL / Sim2Real / Embedded | Built STM32 H7 communication stack, tuned PPO reward functions, and deployed policies on RDK X5. |
| Dexterous Hand Control | Teleoperation / Kinematics | Mapped human hand motion to dexterous-hand joints through kinematics and optimization. |
| 3D LiDAR Mobile Robot | SLAM / Planning / Control | Built localization, global/local planning, simulation, serial communication, and gimbal control pipelines. |
- First Prize, IEEE ICRA 2024 Agilex Sim2Real Challenge
- Second Prize, IEEE ICRA 2025 Sim2Real Challenge
- First Prize, RoboMaster University League Competition, Sentry Robot Group
- First Prize, National Undergraduate Electronic Design Contest, Guangdong Province
- First author, Designing of a Dexterous Hand and Performance Evaluation Based on Teleoperation (IEEE Xplore)
Robots should not only move. They should perceive, reason, and act reliably in the physical world.