Computer Vision Engineer Sports Analytics

exar.live drives the data behind some of the most exciting and innovative experiential and analysis platforms on the market. We offer a unified approach to a wide landscape of sports and entertainment avenues and work with some of the largest broadcasters, clubs, and leagues in the world to unlock the power of 3D analysis and playback. We create powerful tools for use on TV, in club analysis rooms, and at home by fans. Our technologies power the only official Premier League virtual reality game, and have also been used to power the world's first virtual reality broadcast of a football match, as well as allowing users to watch motorsport with more insight and detail than ever before on the participants. The future of what we plan to offer with our technology is near limitless as we step into an exciting new phase of growth. exar.live is part of our parent company, Rezzil, which provides innovative training and performance analysis tools to elite-level sports.

You will design, build, and ship computer-vision systems that extract reliable signals from video (and related sensor streams) and turn them into production-grade sports analytics features. This includes model development, evaluation, optimisation for real-time or near-real-time performance, and robust deployment into live products.

• Build CV/ML pipelines for sports analytics tasks such as detection, segmentation, pose estimation, tracking, action recognition, ball/player tracking, or 3D reconstruction.
• Develop and maintain data pipelines: collection, labelling strategy, quality checks, dataset versioning, and experiment tracking.
• Train, tune, and evaluate models with strong statistical rigour and clear metrics (accuracy, latency, stability, drift).
• Optimise models for deployment (quantisation, pruning, TensorRT/ONNX, batching/streaming, GPU utilisation).
• Collaborate with product, design, and platform engineering to integrate models into user-facing features and services.
• Own model monitoring in production: performance dashboards, alerts, retraining triggers, and incident response.
• Contribute to technical direction: architectural choices, tooling, standards, and best practices.
• Write clear technical documentation and communicate trade-offs to non-specialists.

• Demonstrable sports analytics experience (professional, academic, personal projects, or hobbyist)–e.g., match analysis, player tracking/metrics, event tagging, tactical analysis, or building tools using sports data/video.
• Strong practical experience in computer vision and deep learning, with evidence of shipped systems or robust prototypes.
• Excellent Python skills, plus solid software engineering fundamentals (testing, CI/CD, code review).
• Experience with PyTorch (preferred) or TensorFlow; familiarity with OpenCV and modern CV tooling.
• Strong understanding of CV fundamentals (geometry, camera models, multi-view, filtering) as relevant to the role.
• Experience deploying ML to production (APIs/services, edge or cloud inference, containerisation).
• Comfortable working with GPUs and performance profiling/optimisation.

• Experience in ReactJS/Vite/NPM management
• Familiarity with maintaining Linux systems on the command line.
• Experience in dev-ops (predominantly CI/CD).
• Ansible and/or other automation frameworks
• Video streaming (HLS, DASH, RTMP, SRT, WebRTC)
• Web sockets or Socket.IO
• Computer Vision
• Unity
• Unreal Engine
• Swift and Swift UI
• Electron

• Share options - a chance to participate in the long-term success of Rezzil.
• 25 days\' holiday, plus UK bank holidays.
• Christmas closure - Company shutdown days over the Christmas period.
• Pension scheme in line with UK auto-enrolment.
• Supportive culture - collaborative, inclusive, and focused on sustainable ways of working.
• The chance to work on exciting and innovative projects, either on your own or as part of a group, greenfield or otherwise.