Data Scientist - (Computer Vision)
Role details
Job location
Tech stack
Job description
We are expanding our data science team to include a dedicated computer vision capability. This role is ideal for a candidate who has delivered temporal multi-object tracking and detection systems and understands the full lifecycle from development through to production. You will lead the technical direction of this work, define what we build and how, and play a key role in shaping a growing computer vision team around you.
What You Will Be Doing :
-
Leading the development of multi-object tracking and detection systems from video across multiple sports
-
Defining event taxonomies and annotation strategies for sports-specific training data
-
Building and deploying tracking pipelines, including handling occlusion, re-identification, and camera calibration
-
Designing the output data structures and formats that integrate with our existing GPS and IMU products
-
Collaborating with sports scientists, product owners, and software engineers to translate performance requirements into technical solutions
Requirements
-
Production experience in computer vision, with a focus on multi-object tracking and detection. Experience in sport is a strong bonus
-
Deep understanding of temporal reasoning across video sequences, including re-identification, track association, and handling occlusion
-
Experience with video-native Computer Vision pipelines
-
Understanding of camera geometry, pitch or court calibration, and object localisation from single or broadcast cameras
-
Proficiency in Python and deep learning frameworks
-
Experience with cloud platforms
-
Strong communication skills: you will work closely with non-technical stakeholders including coaches and sports scientists
Nice to Have (But Not Essential):
-
Computer vision experience in sports
-
Familiarity with how tracking and event data is structured and consumed by clubs, analysts, and performance staff
-
Background in biomechanics, pose estimation, or signal processing from IMU/GPS sensors
-
MLOps experience: model versioning, monitoring, and CI/CD for CV pipelines