Research Engineer

The Opportunity

As a Research Engineer, you’ll work on multimodal sensing systems and sensor fusion research for embodied AI and robotics. This role sits at the intersection of robotics, perception, hardware systems, and machine learning, where your work directly shapes how multimodal data is collected, synchronized, fused, and used for downstream VLA training.

You’ll research emerging sensing technologies across RGB-D video, motion capture, IMUs, tactile sensing, audio, and wearable systems, and study how different sensor combinations impact robot learning, policy performance, and generalization. You’ll work closely with hardware, ML, and research teams to design experiments, evaluate new sensing stacks, and help define the next generation of multimodal robotics datasets.

Your work will help shape how frontier labs and leading robotics companies train their models, transforming physical labor markets and economies while contributing to broader research into human embodied intelligence.

What You’ll Do

• Evaluate how different sensor modalities impact VLA training and downstream robotics performance
• Work across RGB-D video, IMUs, motion capture, tactile sensing, audio, and wearable systems data
• Design experiments around synchronization, calibration, fusion, and multimodal alignment
• Prototype quickly and iterate from real-world robotics deployments and research feedback
• Collaborate closely with hardware, ML, and research teams on next-generation sensing systems

What We’re Looking for

• Master’s or PhD in robotics, computer vision, sensing systems, or related fields
• Published research in sensor fusion, perception systems, multimodal datasets, or embodied AI
• Strong technical intuition around hardware systems and robot learning
• Experience with RGB-D video, IMUs, motion capture, tactile sensing, or robotics systems
• Experience with reinforcement learning, real-world robot deployments, and how data impacts downstream policy performance
• Highly curious, execution-oriented, and comfortable operating from first principles