Research Scientist, Spatial AI & Perception
Core
Build perception and geometric reasoning systems for humanoid robots to achieve grounded 3D understanding of the world.
Role type
Research Scientist, Spatial AI & Perception
Builds
Real-time SLAM/state estimation for mobile humanoids; offline 3D reconstruction pipelines for VLM/VLA training data.
Domain
Robotics, Computer Vision, Spatial AI, Large Language/Vision Models (VLM/VLA)
Deliverable
production ML models | product features
Required skills
Real-time SLAM, visual-inertial odometry, state estimation, 3D reconstruction (SfM/MVS), probabilistic state estimation, sensor fusion, geometric foundation models, Lie groups, nonlinear optimization, C++, Python
Preferred skills
PyTorch, JAX, building data pipelines for large model training, VLA/behavior models, top-tier conference publications
Technologies
C++, Python, PyTorch, JAX
Responsibilities
Design and implement real-time SLAM and perception-based state estimation for mobile humanoids; Build offline 3D reconstruction pipelines to generate geometric scene structure; Pioneer research integrating large VLA/VLM models with 3D spatial perception; Bridge classical geometric methods and learned approaches; Write high-quality, maintainable C++ and Python code.
Seniority
Senior, hands-on IC
