Principal Robotics Machine Learning Engineer
Core
Design, train, and deploy end-to-end learned manipulation systems for humanoid robots to perform open-world object manipulation tasks with physical hardware.
Role type
Principal Robotics Machine Learning Engineer (Manipulation)
Builds
Production ML models and data pipelines for the Apollo humanoid robot's manipulation capabilities
Domain
Robotics, Applied AI, Computer Vision, Deep Learning
Deliverable
production ML models
Required skills
CNNs, transformers, Python, C++, PyTorch, 6D pose estimation, camera calibration, point cloud processing, visual-servoing, high-DOF robotic manipulator tuning, data collection strategy, synthetic data generation
Preferred skills
distributed training, MLOps, VR/haptic interfaces, physics engines (IsaacSim, MuJoCo, Drake), logistics automation
Technologies
PyTorch, IsaacSim, MuJoCo, Drake
Responsibilities
Implement and deploy state-of-the-art ML models for open world object manipulation; Drive the data engine for manipulation learning including data collection, curation, and synthetic/real data mix; Drive the entire development cycle from prototyping in simulation to on-robot deployment; Identify and prioritize performance bottlenecks between classical vision issues and model/data issues; Act as a force multiplier by mentoring the next generation of robotics leaders
Seniority
Principal, hands-on IC with mentorship
