Machine Learning at Spotify

Event space and refreshments sponsored by Spotify

Agenda:

6:30pm – 7pm Networking and food

7:00pm – 7:20pm Content-based targeting for ads by Rozmin Daya

7:20pm – 7:40pm Learning a Large-Scale Vocal Similarity Embedding by Aparna Kumar

7:40pm – 8:00pm Driving Insights through Quant & Qual Data by Ye Zhao

8:00pm – 9:00pm Networking

Content-based targeting for ads by Rozmin Daya

Spotify data can be leveraged to identify moods, habits and audiences that advertisers want to reach. This talk describes a Machine Learning system for tagging curated content with multiple labels corresponding to categories of interest.

Learning a Large-Scale Vocal Similarity Embedding by Aparna Kumar

Vocal style and timbre play an important role in musical taste. In this talk we will describe a model for learning a low-dimensional vocal similarity embedding on a large music collection.

Driving Insights through Quant & Qual Data by Ye Zhao

Learn about how the Product Insights team at Spotify uses both quantitative data (AB Tests, statistical analysis) and qualitative data (survey, user interview etc.) to derive actionable insights for both product definition and iterative product development.