Hema Raghavan, Kumo | Databricks Data + AI Summit 2023
summary
Kumo claims to run graph neural networks and AutoML directly on relational data by turning tables into a graph representation and turns tables into a graph to learn embeddings for recommendations. The platform is a 30‑person, 1.5‑year SaaS with guardrails for security; the new integration keeps Delta, Postgres and time‑series data in Databricks with compute pushed to Databricks so customer data never leaves while model training and serving happen in‑place.
AI-written summary · about
chapters (7)
- 0:00Introduction to the talk
- 0:43Lakehouse Apps announcement
- 2:52Platform strategy and ecosystem
- 6:06Kumo's AI on data warehouses
- 8:36Democratizing AI development
- 10:39Opportunities for young developers
- 11:45Call to action and hiring
related talks
52mElevate your data and AI governance with Databricks Data Intelligence Platform
02h 52mKeynote Day 1 - Full
30mGraph Data Science
34mOverview of the Databricks Platform
49mModern Graph Databases
23mAnalystANGLE
49mGenAI for Business on the Modern Data Stack
18mElevating enterprise data through open source LLMs