A physical AI system does not learn from the world in the abstract. It learns from frames, labels, edge cases and the decisions somebody made about what the machine should notice.
Voxel51 and Nebius Developer Community bring that work into a hands-on New York workshop on November 14. Participants will use FiftyOne to examine computer vision datasets and the workflows that turn raw visual information into something a model can learn from.
Robotics and physical AI often arrive through the machine on stage. The harder record sits behind it. A missed object, weak label or unexamined slice of the dataset can travel into the system that acts in a warehouse, vehicle or production environment.
The workshop asks participants to keep the dataset open and work through the tools directly. A laptop is required. Registration requires a company, university or government email address, with exceptions handled by the organizer.
2026-11-14; 1:30-4:30 PM ET.
Kimmel Center for University Life, Room 914, New York.