We provide video analysis tools to professional football clubs and national federations. Our platform processes match and training footage to extract player tracking, event detection and tactical patterns.
We need a CV researcher to improve our player tracking (multi-camera, partial occlusion, broadcast footage) and build new event detection models — tackles, set pieces, transition moments.
You'll have access to a large proprietary dataset of professional match footage and a GPU compute budget.
Publications are welcome but not required. Shipping accurate models that clubs actually use is what matters.
Requirements
– Multi-object tracking experience (ByteTrack, DeepSORT, or similar)
– Broadcast video domain experience (handling compression artefacts, camera cuts)
– Action recognition or event detection models
– PyTorch proficiency
– Interest in football is genuinely a plus — you'll spend a lot of time watching footage