Own what the robot sees: detection, tracking, spatial relations and scene understanding on the sovereign DeepVision tower. Key responsibilities: Train and evaluate perception heads on the frozen DeepVision backbone. Build and curate on-robot perception datasets from real deployments. Own perception latency and accuracy budgets on the on-board compute. Integrate the scene graph with the world model and the control stack. Requirements: Computer vision in production; PyTorch; detection, tracking and segmentation; deploying models to constrained on-board hardware; dataset curation discipline. Salary: Salary on application The Manchester factory comes online Q1 2027. Roles are recruited ahead of that date, with start dates agreed individually. #J-18808-Ljbffr
