Democratizing ML Development to fuel the next generation of Audio Personalization at Spotify

Spotify has expanded its footprint to almost 400 million users across 200 markets around the world in the last 5 years, offering an extensive catalog of 70+ million songs and podcasts. Machine Learning is at the heart of everything at Spotify and enables a personalized user experience to provide billions of fans the opportunity to enjoy and be inspired by the artists on the platform. This is what makes Spotify unique.

In this talk, Divita Vohra, an AI/ML Product Manager on Spotify’s Machine Learning Platform team, will discuss the mission & evolution of the Machine Learning Infrastructure that is the foundation to such user-centric experiences, such as Discover Weekly, Release Radar, and recommendations in your playlist creation screen. She will walk you through a specific personalization use-case, the software architecture that powers it, and the lessons gleaned supporting ML Engineers developing & operationalizing ML solutions at scale.