Award-winning recruiting firm delivering exceptional people, faster.
For CompaniesFor Companies
For CandidatesFor Candidates
Senior Applied Computer Vision Engineer
ApplyApply
Posted about 2 hours ago
Contract
{{salary_package}}
Senior Applied Computer Vision Engineer
Our client, a well-funded sports technology company, is looking for a Senior Applied Computer Vision Engineer to help build and improve video intelligence solutions for sports. This role is focused on applying computer vision and machine learning techniques to real-world sports video workflows. You will work with existing models and pipelines, evaluate performance on new datasets, identify gaps, implement improvements, and partner with engineering teams to deliver production-ready solutions.A central focus is adapting existing models to new video sources (unfamiliar cameras, broadcast styles, and video-quality conditions), where camera calibration and field registration, domain adaptation, and identity association are the hardest and highest-impact problems. The team is looking for someone who combines strong computer vision fundamentals with practical engineering experience and a bias toward execution. The ideal candidate can independently drive initiatives from evaluation and prototyping through deployment and ongoing improvement.
What You’ll Do
Computer Vision Development
- Develop and improve computer vision models for sports video, including player and ball detection, multi-object tracking, event recognition, and identity association (jersey-number recognition / OCR and appearance-based re-identification).
- Build and improve camera calibration and field registration: estimate homographies and map image-space detections to real-world field coordinates across static, moving, and auto-directed cameras.
- Evaluate model performance across different video sources and identify opportunities for improvement.
- Analyze failure modes and implement solutions that improve accuracy, reliability, and robustness.
- Adapt existing models and pipelines to support new sports, camera configurations, and video quality conditions through domain adaptation and transfer learning (e.g., fine-tuning models trained on one league or camera to perform on new ones).
- Partner with data teams on dataset creation, labeling workflows, and quality improvement initiatives.
Production Delivery
- Work closely with software and platform engineers to deploy models into production environments.
- Improve inference performance, scalability, and operational reliability.
- Establish evaluation metrics, testing processes, and quality controls to ensure model performance remains consistent over time.
- Contribute to tooling and processes that make model development and deployment more efficient.
What We’re Looking For
Required Qualifications
- Strong hands-on experience building and improving computer vision systems.
- Proficiency with Python and modern machine learning frameworks such as PyTorch.
- Experience working with video-based computer vision problems, including detection, tracking, event recognition, or identity association.
- Working knowledge of geometric computer vision: camera calibration, homography and projective geometry, and mapping image coordinates to real-world coordinates.
- Experience evaluating model performance, identifying failure modes, and implementing practical improvements.
- Experience adapting models to challenging real-world data where video quality, camera angles, and environmental conditions vary significantly, including domain adaptation / transfer learning across different data distributions.
- Strong software engineering fundamentals and the ability to write maintainable, production-quality code.
- Ability to work independently, prioritize effectively, and drive projects to completion.
Preferred Qualifications
- Experience working with sports video or related domains.
- Experience with large-scale video processing pipelines.
- Familiarity with tools such as FFmpeg and GPU-accelerated video workflows.
- Familiarity with OCR / scene-text recognition (e.g., reading jersey numbers or scoreboard graphics).
- Experience with experiment tracking and model/data versioning (e.g., Weights & Biases, MLflow, lakeFS/DVC).
- Experience deploying machine learning models into production environments.
- Experience with model monitoring, performance tracking, and operational support.
- Experience with human pose estimation (a forward-looking capability for this role).
What Success Looks Like
- Quickly understand existing computer vision systems and establish performance baselines.
- Identify the most important opportunities to improve accuracy, reliability, and scalability, including coordinate and registration accuracy and identity association on new video sources.
- Successfully adapt existing solutions to perform well across new video sources and environments.
- Deliver measurable improvements that can be deployed and maintained in production.
- Operate as a trusted technical leader who can move initiatives forward with minimal oversight.
This search is being conducted on a confidential basis by Fuel Talent on behalf of our client. Full company details will be shared with qualified candidates during the interview process.
