Senior Machine Learning Engineer at Spotify, Sanket Gupta, discusses foundational embeddings for transfer learning in recommender systems. Topics include large-scale recommender system building, transfer learning with user and item embeddings, system evaluation, and MLOps challenges. They explore music recommendation intricacies, user behavior analysis challenges, and balancing real-time recommendations with scalability. The podcast delves into user representations, cross-content embeddings, and maintaining content freshness for optimal user experiences.
