Devoteam
64 Case Studies
A Devoteam Case Study
The Royal Belgian Football Association (RBFA) needed to create a personalized digital experience for its users to keep them engaged with its content, advertisements, and products. They faced the challenge of building a complex, fast, and maintainable recommender system that could overcome issues like data drift. Partnering with Devoteam, the goal was to first develop a system for recommending content tags to users.
Devoteam developed a neural matrix factorisation model using Google's Vertex AI to create high-dimensional embeddings for users and tags, enabling four types of powerful recommendations. The solution was deployed with a fully serverless architecture on Vertex AI Endpoints, App Engine, and Firestore to ensure low latency, meeting an SLO that 95% of requests are served under 100ms. For the RBFA, this resulted in an automated, easily retrainable model that provides a greatly improved personalised user experience and lays the foundation for future ad and product recommendations.
Koen Landsheere
Agile Digital IT Manager