Gasgoo Munich- Gasgoo Embodied Intelligence has learned that Ding Yan, co-CTO of Lumos Robotics, has officially left the company. His social media profile lists him as a researcher at the Zuquan Institute at Fudan University, confirming the career pivot.
This marks a tenure of fewer than 280 days at Lumos Robotics, from his arrival in November 2025 to his departure.
Image source: Screenshot from Xiaohongshu
As an early core driver of the domestic UMI (Unreal Motion Imitation) technology roadmap, Ding holds significant influence in the embodied intelligence data sector. He holds a Ph.D. in computer science from the State University of New York. He previously conducted frontier research in robotics perception and control at overseas institutions, including Ford and Meta. He was also among the first globally to integrate large models into robotics control.
Upon returning to China in 2024, Ding served as a young researcher at the Shanghai Artificial Intelligence Laboratory. He later became co-founder and CTO of Yixing Robotics. During his tenure at the Shanghai AI Lab, he spearheaded the development of several VLA models. This accumulated full-stack experience in embodied large model research.
In late 2025, Yixing Robotics was dissolved due to investors' strategic retrenchment. Ding brought over a dozen core technical team members to Lumos Robotics. There, he led the construction of the UMI data collection system and the technical stack for embodied model training.
Lumos Robotics's FastUMI and FastUMI Pro data collection systems use hardware and algorithm innovations. These boost embodied data collection efficiency fivefold while cutting costs by 80%. This breakthrough breaks the industry's "impossible triangle" of high quality, low cost, and large scale.
The Zuquan Institute at Fudan University, Ding's new affiliation, serves as the university's core platform for technology transfer. It focuses on basic research and industrial implementation in frontier fields like brain-inspired intelligence and embodied intelligence. Research areas include embodied perception, robot control, and trustworthy AI.
