Coco Robotics appoints UCLA professor to head new physical AI lab

Coco Robotics, a company known for its fleet of last-mile delivery robots, aims to extract more insights from the five years of data collected by its machines. Their solution: establishing a physical AI laboratory led by University of California Los Angeles (UCLA) professor Bolei Zhou.

Coco Robotics announced on Tuesday that Zhou has also joined the Los Angeles-based startup as its chief AI scientist.

At its launch in 2020, the company relied on teleoperators to help its robots navigate delivery route obstacles. Coco Robotics co-founder and CEO Zach Rash explained to TechCrunch that the company's long-term goal has always been full autonomy for its last-mile delivery robots to reduce overall delivery expenses. Now, Rash stated, the company has gathered enough data to advance further into automation.

"We've collected millions of miles of data from some of the most complex urban environments, which is incredibly valuable for training practical and dependable real-world AI systems," Rash said. "We've reached a point where our data volume is substantial enough to significantly accelerate research in physical AI."

Rash described the decision to appoint Zhou as head of the initiative as a "straightforward choice." Zhou's research in computer vision and robotics has primarily concentrated on micromobility rather than full-scale vehicles, Rash noted.

Coco Robotics had already been working with Zhou. Both Rash and his co-founder Brad Squicciarini are UCLA graduates and have even donated one of their robots to the university's research laboratory.

"[Zhou] ranks among the world's foremost experts in robot navigation, reinforcement learning, and several other research areas crucial to our work," Rash commented. "He has already demonstrated an ability to attract top researchers he's previously collaborated with to join Coco and help accelerate our progress."

This new research facility operates separately from the robotics startup's collaboration with OpenAI, which permits Coco Robotics to utilize OpenAI's models while granting the AI research lab access to the company's robot-gathered data.

For the present, Coco Robotics intends to use the insights and research from the lab for its internal purposes. Rash mentioned the company has no plans to sell this data to competitors.

Instead, the information will help enhance the company's automation and efficiency, particularly regarding the local models operating on its robots. Rash added that they also plan to share relevant research findings with the cities where they operate to help address obstacles and infrastructure issues that impede their robots' performance.

"The success of this lab will be measured by our ability to deliver higher-quality service at exceptionally low prices," Rash stated. "How can we reduce our costs further? How can we make this more affordable for businesses and consumers? We believe this approach will generate substantial growth within this ecosystem."