Staff, Robotics ML/Data Engineer
Core
Architect end-to-end multimodal data ingestion pipelines and validation systems for robotics training datasets, converting egocentric video, teleoperation recordings, and sensor signals into indexed, queryable, training-ready data.
Role type
Staff, Robotics ML/Data Engineer
Builds
Production-grade multimodal datasets, cross-modal validation systems, and data curation tools for robot learning.
Domain
Robotics, Computer Vision, Time-Series Data Engineering
Deliverable
production ML models
Required skills
Python, PyTorch, video processing pipelines, OpenCV, FFmpeg, Decord, 3D geometry, coordinate transforms, kinematics, URDF, data augmentation, embedding-based retrieval
Preferred skills
Open X-Embodiment, DROID, AgiBot World, EgoDex, SAM-family segmentation, monocular depth, MANO/SMPL, 6-DoF pose tracking, Ray, Apache Spark, synthetic data generation (Omniverse, MuJoCo), Fourier encoding
Responsibilities
Design cross-modal validation systems for video, proprioception, force/haptic signals, and language annotations; Implement spatial, temporal, viewpoint, and sensor-noise augmentation for expert trajectories; Create unified state-action representations across embodiments and coordinate frames; Build tools for dataset querying, visualization, auditing, and annotation workflows; Perform temporal segmentation, metadata extraction, and language annotation of recordings; Use model-failure analyses to develop data curation rules.
Seniority
Staff, hands-on IC
