Coco Robotics taps UCLA professor to lead new physical AI research lab (techcrunch.com)

Coco Robotics has launched a new physical-AI research lab and named UCLA professor Bolei Zhou as chief AI scientist to unlock insights from five years and “millions of miles” of last-mile delivery data collected in dense urban environments. The startup, which initially relied on teleoperators, says it now has enough real-world scale to aggressively pursue autonomy—using the dataset to train more reliable perception, navigation and control systems that reduce delivery costs. The move signals a shift from operator-assisted deployments to research-driven automation aimed at practical, low-cost service delivery. Technically, the lab will focus on problems central to robot navigation, computer vision and reinforcement learning—areas where Zhou has deep expertise in micromobility and recruiting top researchers. The effort is separate from Coco’s OpenAI collaboration (Coco can use OpenAI models while the lab accesses robot-collected data), and the company intends to use the data internally to improve on-device models rather than sell datasets. Coco also plans to share applicable findings with cities to address infrastructure bottlenecks. For the AI/ML community, this represents a rare, large-scale, real-world robotics dataset and an industry-directed research hub that could accelerate applied advances in embodied perception, RL for navigation, and deployment-ready autonomy.