Your mission & challenges 

You’ll join the AI/Perception team to enable our robots to see, understand and interact with dynamic, human-occupied environments. You’ll be responsible for both designing deep-learning perception models and building the full sensor fusion and perception stack for humanoid and collaborative robots.

• Help us architect and build the full perception pipeline: from raw sensor input (cameras, LiDAR, IMU, force/torque, joint sensors) through sensor-fusion, object detection and tracking, 3D reconstruction, scene understanding, up to producing actionable world models and supplying necessary information to downstream modules (navigation, manipulation, interaction).
• Research the state-of-the-art in robot perception and deep learning and bring breaking results from academia to the real robot
• Design, train and finetune neural networks and deploy them both in the cloud and on embedded robotic hardware
• Implement real-time perception software in modern C++ and Python as needed; leveraging classical computer vision and GPU acceleration
• Use ROS2 and DDS frameworks for high data-throughput sensor pipelines, and validate modules both in simulation (e.g., Gazebo, Isaac Sim) and on real robot platforms, ensuring safety and robustness.
• Collaborate with hardware, control, planning and manipulation teams: define perception requirements, integrate sensor hardware, calibrate and optimize systems, develop diagnostics and performance tools, produce documentation and support CI/CD practices.