Case Study: Kaizen Gaming achieves real-time personalized recommendations and 12x faster dataset creation with Databricks

Differentiated and innovative online gaming with data and AI

Kaizen Gaming, one of Europe’s fastest-growing game-tech companies, needed to personalize its high-volume online betting and casino experience but was hampered by legacy data infrastructure. With up to 65,000 concurrent users and thousands of transactions per second, teams couldn’t scale models, collaborate effectively, or analyze minute-by-minute player behavior, and manual CRM reward processes were too slow to retain customers.

By moving to Azure Databricks with Delta Lake and MLflow, Kaizen built scalable clusters, shared notebooks and streaming pipelines to produce features, train models and deploy recommendations quickly. Dataset creation became 12x faster, a sportsbook recommender was moved to production in days and achieved 75–85% precision on the top 100 events, and near–real-time bonuses, better targeting and much higher ML productivity now drive improved engagement and revenue.

Kaizen Gaming

Constantinos Liapis

Head of Applied AI