Inside SoundCloud’s analytics: trusted data layer for billions of daily events

At SoundCloud, data drives everything: recommendations, search, personalization. So how do you put it in front of agents and every analyst without losing trust in it? Brian Doherty from SoundCloud walks through how they did it.

🔒

What you’ll learn

• ✓
  How to keep data trusted across thousands of tables

  How SoundCloud’s numbers stay dependable even as AI opens querying to people unfamiliar with the tables underneath.

• ✓
  What “ChatGPT for your data” really means at petabyte scale

  The gap between wanting a simple answer and pulling it from petabytes across 100,000+ tables.

• ✓
  How to build a semantic layer that scales

  SoundCloud shows how they built a semantic layer flexible enough for whatever tool comes next.

Presented by

Brian Doherty

Senior Principal Data Analyst, SoundCloud

Jakub Jurových

Founder & CEO, Deepnote