Building AI ready applications
summary
Shows a working pipeline that keeps data in Azure SQL DB and augments it with Azure OpenAI and Azure AI Search: precomputes embeddings for 8 million reviews, stores vectors as varbinary, applies relational filters (city, >4 stars, 5 km geo radius, regex) then runs a single similarity call — one embedding call at query time — to find results despite spelling errors. Demonstrates stored procedures calling language and content safety endpoints to redact PII and return confidence scores, and a LangChain SQL agent that issues NL-to-SQL queries and runs them.
AI-written summary · about
chapters (9)
- 0:00AI ready applications
- 1:53Intelligent apps on SQL
- 6:09OpenAI from SQL procedures
- 14:20Hybrid search SQL + AI
- 26:37SQL with Azure AI Search
- 35:49Content moderation AI
- 37:59AI agents with LangChain
- 43:00Productionizing AI apps
- 45:48Q&A
related talks
01h 16mAzure SQL ❤️ Open AI
46mPower AI apps and develop rich experiences with Azure SQL Database
44mPostgres and OpenAI - a match made in Azure
44mSupercharge your data with Azure AI Search and OpenAI
46mAdvanced RAG with Azure AI Search - top questions answered
27m6 things you can do with azure_ai & PostgreSQL on Azure
56mDeveloping AI Applications with SQL AI: Ground to Cloud to Fabric
19mAI Driven Data Enrichment with Azure Cognitive Search