Saavn, India's largest music streaming service, has grown to 80m subscribers with the help of MongoDB, leveraging its flexibility and scalability to support rapid growth. By using MongoDB as a system of record for subscriber data, devices, and user activity, Saavn was able to build a sophisticated analytics pipeline that powers personalized push notifications, driving a 3x increase in user engagement. The service's complex data processing pipeline uses Apache Kafka, Storm, and MongoDB to compute relevant playlists based on user preferences, activity, device, location, and more. With the system sending 30m notifications per day, Saavn has scaled its use of artificial intelligence to predict user interests, storing machine learning models in MongoDB for real-time serving. The company's experience with MongoDB has also led to improvements in performance, schema design, and system monitoring through the use of WiredTiger storage engine, TTL indexes, shard key selection, and MongoDB Cloud Manager.