Computer Vision Engineer, Body Sensor Networks

About the role

Lead camera-based perception for motion capture, building deep learning and computer vision pipelines that convert images into human pose estimates. The role works alongside sensor fusion experts, biomechanists and mathematicians and can be hybrid or remote.

What you will do

• Choose, adapt and tune deep learning models for 2D and 3D pose estimation from cameras
• Create camera models and calibration routines for varied camera and lens setups
• Assess motion capture output and build methods to handle vision failure modes
• Construct training and evaluation pipelines from camera and optical capture recordings
• Generate synthetic data when real recordings are insufficient
• Follow computer vision research and judge which new methods suit products

What they look for

• MSc or PhD in computer science, electrical engineering, robotics or similar, plus relevant experience
• Hands-on training and tuning of pose estimation models validated on public benchmarks
• Understanding of camera and lens models, wide field-of-view optics, calibration and multi-view geometry
• Familiarity with pose output representations such as keypoint sets and SMPL-family parameters
• Experience with markerless tools such as OpenCV, Sapiens and SAM3D
• Proficiency in Python and working knowledge of C++ for model integration

Nice to have

• Synthetic data generation and rendering pipelines
• Optimising models for real-time use with ONNX, TensorRT or quantization
• Building datasets and benchmarks for human motion