Jason Osborn of Intel Robotics talks with Dave Coleman, PickNik Robotics founder and chief product officer, at Automate 2026 in Chicago about robot control, compute choices, safety, and a shelf-stocking demo.
Coleman says PickNik builds hardware-agnostic control software for robot arms, supporting different brands and cameras. The company works with organizations tackling difficult automation tasks, including mobile robots that combine wheels and an arm. He describes MoveIt Pro and hybrid autonomy as part of PickNik’s approach.
Coleman points to bathroom cleaning and food-processing facility cleaning as current applications, with data centers and car washes also in development. PickNik combines segmentation, vision-language, and language models with robot control. He says these systems can address tasks beyond conventional industrial automation.
Coleman says the right compute depends on each application’s balance of CPU, GPU, and NPU needs, and notes PickNik has often used Intel NUCs. For safety-critical work around people, he argues that AI models should be checked and supplemented with established planners, controllers, and collision checking. He says CPU-based processing can support predictable behavior and help explain why a robot acted as it did.
Coleman describes a mobile robot project for placing groceries or convenience-store products on shelves. The system identifies and segments items in a mixed container, then plans picks that preserve the products’ intended orientation. Osborn says the demo combines a vision-language model and segmentation with conventional planning, control, collision checking, and safety systems, shown through a live physics model.
Briefs summarise talks and interviews; timestamps open the video at that moment. Claims in a talk are the speaker's own.
Dave Coleman, founder and chief product officer at PickNik Robotics, discusses robotics software and the UX of MoveIt Pro at a Northwest Robotics Alliance talk.
