Semantic search in prod w/ Notion, Cursor & turbopuffer

​Panel Discussion

​Moderated by Braintrust co-founder and CEO Ankur Goyal

​Notion will show how they built AI-powered search across millions of workspaces. Cursor will explain how they make code context actually work. turbopuffer will discuss the infrastructure challenges of serving semantic search at scale. Real production gotchas, architecture decisions, lessons learned from billions of queries and petabytes stored. 

​Who should attend:

• ​Engineers building search, RAG, or agentic AI features

• ​Developers working on code intelligence or content discovery

• ​Technical leads evaluating semantic search approaches

• ​Anyone shipping AI-powered products who wants to learn from teams doing it at scale

​What you'll learn:

• ​Why Notion chose their architecture over others. 

• ​Cursor's approach to understanding and retrieving code context

• ​Infrastructure patterns for serving semantic search reliably

• ​Production challenges you won't read about in tutorials