Summary
Xiao Liu is a Senior Researcher specializing in embodied AI and robot intelligence, currently leading research at Samsung Research America’s Robot Intelligence Lab. With nine years of experience spanning academia and industry, Xiao developed diffusion-based action policies, multimodal latent-state estimators, and differentiable Ensemble Kalman Filters to enable long-horizon dexterous manipulation and imitation learning. Recent work at Honda Research Institute focused on scene-graph assistant generation and context reasoning that bridge large vision models and LLMs for teleoperated avatar robots. Xiao’s background blends mechanical engineering and a PhD in computer science, and he has validated methods on both rigid and soft robots as well as on real-world multi-object tracking and monocular depth tasks. Known for integrating image-generation ideas (e.g., ControlNet-style conditioning) into control policy design, he brings a rare cross-domain fluency in perception, planning, and control. Based in Tempe, Arizona, Xiao combines rigorous algorithmic priors with practical system-building to push robot autonomy toward everyday manipulation.
9 years of coding experience
8 years of employment as a software developer
Doctor of Philosophy - PhD Computer Science, Doctor of Philosophy - PhD Computer Science at Arizona State University
Master's degree Mechanical Engineering, Master's degree Mechanical Engineering at Case Western Reserve University
Bachelor of Engineering (B.Eng.) Mechanical Engineering, Bachelor of Engineering (B.Eng.) Mechanical Engineering at Southwest Jiaotong University
English, Chinese