AI Object Pose Estimation for Manipulation Engineer

About the role

Responsibilities

• Design pose estimation models handling occlusion, reflective surfaces, and cluttered multi-object scenarios
• Implement category-level pose estimation generalizing across object instances without per-object CAD models
• Build simulation-to-real training data pipeline for efficient new object category coverage
• Develop pose uncertainty estimation enabling the manipulation controller to request re-estimation when needed
• Validate pose accuracy against robot grasp success rates as the primary production metric
• Optimize inference for real-time pose estimation on robot perception hardware

Requirements

• 6+ years in computer vision engineering with 3D perception, pose estimation, or robotic vision experience
• Deep expertise in 6DoF object pose estimation and the relevant model architectures
• Strong Python and PyTorch skills with 3D data processing experience
• Familiarity with sim-to-real transfer techniques and synthetic data generation for robotics
• Track record improving grasp success rates through perception system improvements

Benefits

• Fully remote with quarterly visits to robotics lab for physical system testing
• Top-of-market compensation for robotics perception expertise
• Comprehensive health benefits
• $5,000 annual equipment and conference budget
• Equity participation

Keep exploring related roles

Use these topic pages to discover more roles in the same hiring cluster.