Integrating Serverless LLMs with Data Platforms for Natural Language Analytics
summary
Shows how Snowflake Cortex exposes serverless LLM SQL functions and Python/REST APIs that run next to data, demoing Streamlit apps, Slack integration, and a PNG invoice parse that extracts text, chunks it into vectors, and serves semantic+keyword search without any external vector DB via vector columns (no external vector DB). Describes fine-tuned smaller models to cut token cost, batch inferencing, Cortex Guard output safety, Cortex agents for orchestration, and claims 92% text-to-SQL accuracy with a 30-day trial to reproduce examples.
AI-written summary · about
chapters (5)
- 0:00Introduction and resources
- 3:08Easy efficient trusted platform
- 11:56Demo with Snowflake notebook
- 22:33Cortex playground and agents
- 30:34Governance and security summary
related talks
36mImplementing Natural Language Analytics with Snowflake Cortex
49mLLM inference and fine-tuning on secure enterprise data
28mAutomating Database Operations with Cortex Code
29mBuilding End to End Gen AI Powered Analytics: From Data to Production
31mAI-driven DataOps: Building data products with a taste of the future
17mPrepare to master common generative AI use cases with ease
15mFrom Air Quality to Aircraft and Automobiles, Data Is Everywhere
27mUsing GenAI with Audio Data for Driving Better Customer Experience