Andrew Zane, Unstructured | Databricks Data + AI Summit 2024
summary
Company positions itself as ETL for RAG, claiming to ingest and normalize unstructured documents into JSON and chunked metadata for vector stores. It handles 35 file types, connects to top vector DBs (Pinecone, Weaviate, Milvus, MongoDB, etc.), and integrates with Databricks Unity Catalog for downstream search. Open source package shows ~10 million downloads and ~1.2M/month; team ~55, $65M raised; serverless and an enterprise GUI rollouts are imminent.
AI-written summary · about
chapters (7)
- 0:00Introduction and branding
- 1:36Defining Unstructured
- 3:40RAG and enterprise challenges
- 6:14Partnerships and community
- 8:09The evolving AI landscape
- 10:36Future potential of AI
- 12:43Future products and wrap up
related talks
15mUnlock the Power of Your Data for Intelligent AI Applications
34mOverview of the Databricks Platform
03mDay 2 Opening Remarks
29mStructuring Data for Scalable AI Applications
05mDatabricks Product Announcements
05mInterview
32mRevolutionizing ETL for RAG with
58mAI-Enhanced Big Data: Integrating Private LLMs and Vector Databases