Hyperight

Machine Learning for Personalized Music Recommendations at Scale – Rishabh Mehrotra, Spotify

video
play-rounded-fill

Surfacing relevant content from among millions of candidates to users in real time is a challenging task addressed by recommender systems. Modern platforms have customers not only on the demand side (e.g. users), but also on the supply side (e.g. retailer, artists), and face an interesting problem of optimizing their models not only for user satisfaction but also for supplier preferences, and visibility. We discuss a number of ML problems which need to be addressed when developing a recommendation framework powering multistakeholder marketplaces

Key Takeaways

  • Aspects of recommendations across multiple features & products
  • Multi-objective ML methods for multi-stakeholder recommendations
  • User & content understanding for better decisioning
  • Insights from large scale experimentation and deployment of music recommender systems

Add comment