Senior Computer Vision Engineer

About the job

The organization has an exciting opportunity for an experienced and dedicated Senior Computer Vision Engineer to join our team!

The Senior Computer Vision Engineer will innovate, design, develop, and deploy real-time computer vision systems for elite sport. This deeply technical role will contribute to officiating support technology and offers the chance to join a world-first innovation project from the ground up-conducting research and rapid development to deliver optical tracking products.

The role spans object detection and tracking, multi-camera systems, 3D geometry, low-latency models, production-grade deployment, and MLOps.

Key Responsibilities

• Design and implement real-time computer vision algorithms for sport in a multi-camera, latency constrained environment.
• Build robust 3D reconstruction pipelines for problems including ball and player tracking under strong occlusion and noise.
• Build, train and evaluate state-of-the-art deep learning models for problems such as object detection, segmentation, action recognition, event detection and pose estimation.
• Conduct rapid experimentation to quickly validate ideas.
• Contribute to the development of CV/ML solutions across the entire product lifecycle, from ideation/experimentation to POC/MVP to production.

Production Deployment

• Package and deploy CV models to edge and cloud compute.
• Monitor solutions and deploy iterative improvements based on observed performance and stakeholder feedback.
• Write maintainable, well-tested, production ready code, and work closely with software engineers and architects to productionise successful experiments and proof of concepts.
• Design scalable solutions that are suitable for the resource constraints.

Collaboration

• Collaborate with a team of CV engineers, AI engineers, software developers and data analysts, and engage in weekly code/algorithm review meetings.

About You

• Bachelor’s or Master’s degree in Computer Science, Engineering, or a related field.
• Publications or contributions in Computer Vision/Machine Learning research or open source projects are highly regarded.
• 4+ years of industry or research experience in computer vision (prior experience in sport highly regarded but not required).
• Strong experience with modern deep learning models for computer vision (e.g. CNNs, Vision Transformers, VAEs, GANs etc.)
• Experience with common computer vision paradigms including object detection (e.g. YOLO, SSD models), multi-target object tracking (e.g. SORT, DeepSORT), action recognition and pose estimation.
• Experience with multi-camera systems, including a familiarity with projective geometry.
• Proven experience developing computer vision solutions in real-time or latency constrained environments.
• Expert level Python programming including strong experience with common CV/ML and scientific libraries, such as PyTorch, OpenCV, NumPy, SciPy, scikit-learn etc.
• Demonstrated experience working across the entire CV model life-cycle, from rapid experimentation to deployment and monitoring using best practice MLOps.
• Experience working with large volumes of visual data in a real-time context.
• Familiarity with software and MLOps best practices including version control, containerisation, and model and data traceability.
• Familiarity with cloud platforms a plus (e.g. GCP).
• Excellent communications skills and an ability to communicate with technical and non-technical stakeholders.