ScyllaDB and Scala: Scaling Up Machine Learning Experimentation with a NoSQL Database

Tubi, an ad-supported video on demand (AVOD) platform, has successfully leveraged a NoSQL database, ScyllaDB, and the Scala programming language to enhance its machine learning and personalization infrastructure. Initially using a monolithic architecture with Apache Spark and Redis, Tubi faced challenges in scalability and efficiency, particularly in running A/B experiments, which required extensive manual processes and collaboration between machine learning engineers and backend teams. The transition to a new architecture, featuring a Ranking Service and an experiment engine named Popper, streamlined these processes by embedding ScyllaDB and using AWS Kinesis and Apache Airflow for automatic model deployment. This new system, based on Scala/Akka applications, reduced latency, improved maintainability, and allowed engineers to conduct experiments without code changes or manual intervention, thereby optimizing machine learning operations and enhancing user experience.

Use this post, company, and trend context to find content marketing opportunities, perform competitive analysis, or address product feature gaps via the Plushcap MCP server or the Plushcap API.