Esports
Master Thesis Intern
Pre-career
Stockholm, Sweden
in office
Internship
Digital, Product & TechEngineering & DevelopmentInnovation & SportsTech
Job description
About the Role
The Computer Vision and Machine Learning team at EA SPORTS TRACAB in Stockholm offers master's thesis projects focused on capturing, reconstructing, and understanding sports through computer vision and machine learning. Students will work closely with engineers on challenging, open-ended research problems using real-world sports data. The projects are expected to run from February to June 2027, with the precise scope and methodology agreed with the selected student and adapted to their background, interests, and research ideas.
Key Responsibilities
- Work closely with engineers on challenging, open-ended research problems using real-world sports data.
- Investigate whether recent machine learning and foundation-model approaches can automatically identify, localise, and explain potentially unreliable segments in sports tracking data
- And develop a rigorous evaluation framework comparing modern approaches with strong statistical and machine learning baselines.
- Explore systems that provide more than a single quality score, offering information about when data should be investigated and what may have occurred.
- Investigate modern machine learning approaches for automatically recognising football events from visual data, player and ball tracking data, or combinations of the two.
- Work with real football footage, accurate tracking data, and corresponding event annotations to explore vision-language models and models that reason over multivariate spatiotemporal player and ball trajectories.
- Compare different representations in terms of accuracy, generalisation, and the amount of labelled data required.
- Investigate state-of-the-art 3D human reconstruction methods using multi-view footage from close-contact sports, establishing capabilities and limitations and improving reconstruction accuracy and temporal consistency during severe occlusion and physical interaction.
- Investigate how learning-based techniques can complement geometric calibration and numerical optimisation, developing and evaluating methods that extract useful scene information and incorporate it into a principled calibration framework.
- Investigate machine learning methods that combine visual evidence, temporal modelling, and physical constraints to produce accurate and robust 3D ball trajectories, with attention to uncertain or missing observations and uncertainty representation.
- Investigate computer vision and machine learning methods for estimating the position and orientation of the ball throughout play in American football using multi-view footage, exploiting spatial context, temporal information
- And observations from multiple cameras to maintain reliable estimates through occlusion.
Required Experience & Qualifications
- Current master's student undertaking a thesis project.
- Available to undertake the project from February to June 2027.
Preferred Skills
- Interest in foundation models, anomaly detection, time-series learning, and explainable AI.
- Interest in vision-language models, video understanding, temporal modelling, and multimodal learning.
- Interest in cutting-edge computer vision, human motion understanding, multi-view learning, and 3D reconstruction.
- Interest in combining modern machine learning with multi-view geometry, mathematical modelling, and optimisation.
- Interest in computer vision, sequence modelling, probabilistic inference, and physics-informed machine learning.
Create your account to access job details
Unlock full job descriptions, apply in seconds, and get access to roles matched to your experience and interests.
