Case Study: Rave achieves 70% lower operating costs and scalable AI-powered DJ mashups with Google Cloud Platform

Rave Keeping the party going for listeners

Rave is a Canadian social app developer that creates shared music and video experiences on phones and tablets. When users began running long, DJ-like sessions, Rave built RaveDJ—an AI-powered service that analyzes song structure and makes real-time mashups and mixes—but needed a scalable platform to handle intensive audio/video processing and model training.

Rave moved RaveDJ to Google Cloud Platform, using TensorFlow and Cloud ML Engine on Google Kubernetes Engine, Cloud Storage, Stackdriver, Pub/Sub and Compute Engine with preemptible instances. The migration enabled fast model training and reliable real-time processing while cutting server costs by 70%, reducing content delivery costs by 66%, minimizing storage budgets, and producing more than 2 million song mashups.

Rave

Garrett Everding

Chief Technology Officer