Joel Lee

Hello everyone!
I’m a Robotics Research Engineer working on AI systems that help robots act more safely and reliably in the physical world. My goal is to make robots easier for people to use in everyday life by combining Robotics with LLMs (or VLMs).
My research bridges the gap between academic robotics and real-world applications through two key approaches: (1) enhancing physical reliability using simulation-to-real transfer and (2) improving contextual understanding via Large Language Models. It enables robots to perform complex manipulation tasks while maintaining safety and user interpretability (i.e., commonsense).
I completed my M.S. in Artificial Intelligence at Korea University under the supervision of Sungjoon Choi. During my academic journey, I also gained valuable experience in reinforcement learning for robot navigation at KAIST MDI Lab, where I worked on various robotics and computer vision projects.
Research Topic:
- Physics-based Robot Control and Simulation
- LLM-driven Robotics (e.g, VLA)
- Multimodal Perception and Reasoning (Vision, Language, Audio)
- Human-Robot Interaction and Preference Learning