Roboticist. Working on robotics. machine learning & reinforcement learning. embodied AI.
Programmer & Roboticist
I’m Jun Zeng (曾俊). My work spans robotics, artificial intelligence and autonomous systems — optimal control, nonlinear optimization, motion planning, trajectory generation, reinforcement learning and machine learning. I build algorithms at the intersection of decision-making, learning and physical systems, so that robots and autonomous vehicles can operate safely and intelligently in complex real-world environments.
I’m currently the Chief Technology Officer of Cyan Robotics in Shanghai, leading Amoo, an embodied companion robot, from research prototype to consumer product. I received my Ph.D. from the Hybrid Robotics Group at UC Berkeley under Prof. Koushil Sreenath, where I worked on model-based optimization, control and learning for aerial and legged robots.
I’m always glad to have technical discussions and to build new connections — feel free to reach me through the links below.
Own the technology strategy, roadmap and system architecture for Amoo, an embodied companion robot. Lead a multidisciplinary team across the full robotics and AI stack — hardware-software integration, robot control and autonomous behaviors through foundation models and LLM-based intelligence.
Trajectory generation and large driving models for autonomous vehicles: unified reactive motion planning and control, non-convex trajectory generation solvers, and constraint enforcement inside large behavior models across pretraining, post-training and deployment.
Advisor: Prof. Koushil Sreenath
Field: Control, Robotics, Machine Learning
Reviewer for 18 journals and 10 conferences in robotics and control, including T-RO, TAC, RA-L, IJRR, ICRA, IROS and CDC. — Full list →
No posts yet
Visit blog