Kai Lu is a postdoctoral researcher at the Oxford Robotics Institute, working with Prof. Ioannis Havoutis on perceptive contact-rich loco-manipulation. His research combines generalizable policy learning, dynamic robot interaction, and multimodal models. He received his Ph.D. in Computer Science from the University of Oxford in 2025, supervised by Prof. Andrew Markham and advised by Prof. Niki Trigoni FREng, with a thesis on embodied visual control policy learning for mobile manipulation. During his Ph.D., he interned at Mitsubishi Electric Research Labs (MERL) and collaborated with Meta AI on multimodal robotic action generation. His work has been published at NeurIPS, RA-L, ICRA, and IROS, and featured in Oxford CS Department News and British Science Week 2025. He received his B.E. in Automation from Tsinghua University in 2020 as an Outstanding Graduate, and previously interned with Prof. Kris Hauser at Duke and UIUC on robotic manipulation.
Most Recent Publications
InteLiPlan: an interactive lightweight LLM-based planner for domestic robot autonomy
InteLiPlan: an interactive lightweight LLM-based planner for domestic robot autonomy
