Senior Applied Computer Vision Engineer (Sports Analytics)

TL;DR

Senior Applied Computer Vision Engineer (Computer Vision/PyTorch): Designing and scaling computer vision systems for sports video analytics with an accent on video-based detection, tracking, and geometric camera calibration. Focus on transforming sports video into actionable insights through homography estimation, identity association, and production-grade ML pipelines.

Location: Remote (European Residence required)

Company

Engineering team specializing in high-performance solutions for the sports analytics industry.

What you will do

• Develop and optimize computer vision models for player and ball detection, tracking, and event recognition in sports video.
• Build camera calibration, homography, and field-registration solutions to map image coordinates into normalized field coordinates.
• Analyze existing pipelines to identify bottlenecks and implement practical improvements to increase accuracy and reliability.
• Design end-to-end experiments covering data acquisition, augmentation, model training, fine-tuning, and deployment.
• Collaborate with data teams on labeling workflows, dataset quality, and human-in-the-loop improvement cycles.
• Partner with DevOps and platform engineers to ensure inference performance, scalability, and operational reliability in production.

Requirements

• Strong experience building production-grade computer vision systems.
• Proficiency with Python and PyTorch.
• Expertise in video-based CV problems: object detection, multi-object tracking, and identity association.
• Deep knowledge of geometric computer vision: camera calibration, homography, and projective geometry.
• Experience adapting models to challenging real-world data with varying video quality and camera configurations.
• Must be a resident of Europe.

Nice to have

• Experience with sports video, broadcast analytics, or American football.
• Familiarity with FFmpeg, GPU-accelerated video workflows, and inference optimization.
• Experience with OCR and appearance-based re-identification (jersey-number recognition).
• Proficiency with experiment tracking tools like MLflow, Weights & Biases, DVC, or lakeFS.
• Experience with MLOps practices and deploying ML models into production environments.

Culture & Benefits

• Competitive compensation with paid vacation and sick leave.
• Ultra-flexible remote-first working conditions and flexible hours.
• Generous office equipment allowance or provision of a coworking desk.
• Collaborative startup environment with a globally diverse team of top engineering talent.