Principal Robotics Machine Learning Engineer

The Principal Manipulation Engineer is a core contributor to our robot’s ability to interact with the world with human-like world model context, dynamic perception, and interactive behavior. This role is responsible for the end-to-end lifecycle of learned manipulation, from data engine design through training to on-robot deployment and is expected to diagnose performance bottlenecks. This requires deep fluency in both classical computer vision and modern deep learning to unlock the full potential of state-of-the-art humanoid robot hardware.

This role will bridge the gap between cutting-edge research and scalable, reliable production software. We move at startup pace; intense and focused, but sustainable. We operate with high ownership, fast feedback, and low bureaucracy.

• Implement and deploy state-of-the-art ML models/algorithms to achieve ambitious, world-class performance on open world object manipulation tasks with physical hardware.
• Drive the data engine for manipulation learning: data collection strategy, curation, annotation, and synthetic/real data mix, in close partnership with data infrastructure teams.
• Drive the entire development cycle, from prototyping in simulation to robustly transferring and fine-tuning models on the robot.
• Identify and prioritize performance bottlenecks, distinguishing between root causes traceable to classical vision (calibration, geometry, sensor fusion) vs those rooted in sub-optimal model/data, to prioritize fixes according to customer-expected levels of reliability.
• Act as a force multiplier across the organization. Beyond code reviews, foster a culture of technical rigor, setting the bar for architectural excellence and mentoring the next generation of robotics leaders.

SKILLS AND REQUIREMENTS

Technical Skills (Must-Have)

…