We work on the dexterity problem for humanoids

🎯 Role Details

• Design and deploy real-time SLAM and state-estimation systems for humanoid and mobile robotic platforms
• Build multi-sensor fusion pipelines (RGB-D, stereo, LiDAR, IMU, tactile)
• Develop neural scene representations including NeRFs, neural occupancy grids, and signed distance fields
• Bridge classical geometry-based perception with learning-based spatial models
• Enable long-horizon 3D world modeling for manipulation and interaction
• Integrate perception outputs into VLA (Vision-Language-Action) systems
• Improve robustness under occlusion, motion blur, lighting changes, and sensor noise
• Support sim-to-real transfer using synthetic data and domain randomization
• Collaborate closely with robotics, controls, and learning teams

⚡ Requirements

• MS or PhD in Robotics, Computer Vision, Computer Science, or equivalent experience
• Strong background in SLAM, state estimation, and sensor fusion
• Deep understanding of 3D geometry, multi-view geometry, and camera models
• Experience with NeRFs, neural occupancy, or implicit scene representations
• Familiarity with Vision-Language-Action (VLA) or embodied AI systems
• Proficiency in Python and C++ on Linux
• Hands-on experience with real robot perception stacks
• Comfortable working with large datasets and long-running experiments
• Self-driven, systems-oriented, and excited about real-world deployment

Bonus

• Experience with tactile or event-based sensors
• Background in Gaussian splatting, neural SDFs, or hybrid map representations
• Experience integrating perception into manipulation pipelines
• Familiarity with Isaac Sim, MuJoCo, or photorealistic simulators

Robotics hardware + data collection + AI