Robot Learning Engineer
Core
Define camera and data capture specifications for warehouse robotics, build evaluation harnesses, and post-train open robot-learning models to enable grocery manipulation tasks.
Role type
Robot Learning Engineer (Computer Vision & Data Pipeline)
Builds
Large-scale grocery manipulation datasets, evaluation harnesses, and post-trained robot-learning models
Domain
Robotics, Computer Vision, Warehouse Automation
Deliverable
production ML models | dashboards & analysis
Required skills
PyTorch, computer vision (3D geometry), camera calibration, pose estimation, SLAM, ROS 2, hand-pose estimation, egocentric video processing, multi-camera rig synchronization, dataset failure mode diagnosis
Preferred skills
debugging extrinsic calibration, reading and reproducing published research, operating independently as sole ML specialist
Technologies
PyTorch, ROS 2, cloud computing infrastructure
Responsibilities
Define camera, mounting, and calibration specifications for warehouse data capture; own hand-pose ground truth measurement and address occlusion challenges; build an evaluation harness to determine suitable recorded hours for training; post-train open robot-learning models and run initial task evaluations; read and reproduce published research to guide data-capture investments; direct AI agents for harness, data-pipeline, and reproduction work
Seniority
Individual Contributor, hands-on specialist
