Research Interests

I build algorithmic frameworks for Embodied Intelligence, spanning agentic
AI, reinforcement learning, and physics-aware world models.
In humanoid robotics, I work on two complementary levels to bridge
cognition and control:

 • Brain – Agentic AI for task reasoning, long-horizon planning,
    and tool use in bimanual manipulation.

 • Cerebellum – Generalizable low-level skills for bimanual manipulation
    and mobile grasping, driven by physics-aware world models.

Master's students looking for a thesis topic can find my and
my colleagues' offers here.

For more information, please visit my personal website:

https://kchu.github.io/

Publications