Microsoft Events hosts an expert-led Ignite session featuring speakers like Aditya Badramraju and Anna Hoffman, guiding attendees on best practices for scalable, secure AI app development using Azure SQL Database Hyperscale.

Building Scalable AI Apps with Azure SQL Database Hyperscale

Session Overview

This advanced session (Microsoft Ignite 2025, BRK126) addresses modern strategies for developing AI-driven applications that leverage Azure SQL Database Hyperscale for secure scalability.

Key Topics Covered

  • Native Vector Indexes: Achieve efficient semantic search in SQL databases by implementing vector indexing for AI-powered search and retrieval processes.
  • Read Scale-Out for RAG: Support low-latency retrieval-augmented generation (RAG) at scale by distributing read workloads.
  • Secure Model Invocation with Microsoft Foundry: Invoking AI models securely from T-SQL, enabling integration with the AI model of your choice within database workflows.
  • Continuous Priming: Enhance performance and readiness in Hyperscale environments with continuous data preparation and priming.
  • Cost Management Strategies: Utilize serverless and elastic pool features in Azure SQL Database to optimize costs for dynamic workloads.
  • Modernization via Microservices and Cloud Hosting: Migrate and modernize legacy apps with Azure microservices and cloud-native hosting.
  • AI-Driven Automation: Streamline document processing and knowledge graph creation using embedded AI within SQL workflows.
  • Embeddings for Semantic Search: Integrate vector embeddings for advanced AI queries directly within your SQL data estate.

Real-World Insights

Global technology company Hexagon and investment firm BlackRock present their practical experiences integrating Hyperscale features into AI-powered solutions. Learn about architectural decisions, observed benefits, and implementation challenges.

Best Practices & Demos

  • Deep dives into Hyperscale scalability and architecture
  • Live demonstrations on semantic search, embeddings, and continuous priming
  • Guidance on securely calling AI models from Microsoft Foundry
  • Tips for cost management and workload optimization

Further Resources

Speakers

  • Aditya Badramraju
  • Dmitry Borodin
  • Britt Ewen
  • Anna Hoffman
  • Ross Jenkins