Case Study: Brave achieves faster ML training and 50% lower infrastructure costs with Rackspace Technology MLOps on AWS

Brave Software improves its web browser using machine learning on AWS

Brave Software, the privacy-focused browser that rewards users with Basic Attention Token (BAT) and supports advertisers and publishers, needed to scale its machine-learning model that classifies websites for advertising. With more than 30 million monthly active users and over 1 million verified publishers, Brave required a robust, auditable way to deploy, manage, and update models so it could reliably serve ads while preserving user privacy.

Brave partnered with Onica (a Rackspace Technology company) and AWS Jumpstart to implement Onica’s MLOps Foundations using AWS services including SageMaker, CodePipeline, Lambda, CodeCommit, Systems Manager, and S3. The automated CI/CD pipeline and model optimizations reduced model training from several days to six hours, cut infrastructure costs by 50%, sped deployments to hours, added QA/approval workflows, and improved responsiveness to advertisers — all delivered ahead of schedule.

Brave

Jimmy Secretan

VP of Services and Operations