Microsoft Events presents a hands-on Ignite 2025 session on optimizing Azure Storage for secure, scalable AI workloads. Speakers Vamshi Kommineni, Natalie Mao, and Saurabh Sensharma guide viewers through advanced techniques and integrations.

Running AI on Azure Storage: Fast, Secure, and Scalable

Overview

This advanced session from Microsoft Ignite 2025 explores how Azure Storage, especially Blob Storage, powers high-performance AI workloads. The presentation showcases integration strategies and best practices for scaling OpenAI models, maintaining GPU productivity, and building secure data pipelines for enterprise AI applications.

Key Topics Covered

1. Azure Storage for AI Workloads

• Azure Blob Storage: Scales to meet demanding AI data requirements.
• Azure Container Storage & Blobfuse2: Ensure GPUs remain utilized and speed up access.

2. Integration with AI Frameworks

• Ray/KAITO Integration for AKS: Simplifies running distributed AI jobs on Azure Kubernetes Service.
• LangChain on Azure: Easy method for AI app development and retrieval augmented generation (RAG).

3. AI Training and Data Preparation

• Deep Dive into workflow for training and fine-tuning AI models on Azure.
• Data Preparation Steps (incl. RAG): Convert enterprise data to formats ready for AI consumption.

4. Security and Performance

• Network & Data Security practices for AI environments.
• Unified management through Azure Storage Center.
• Methods for maximizing developer velocity and infrastructure reliability.

5. Demonstrations

• Live demos of Azure Storage Center and integrating storage with Microsoft AI services.

6. Best Practices

• Tips for performance tuning, secure storage access, and scalable AI solution architecture.
• Guidance on blending open-source frameworks (Ray, KAITO, LangChain) with managed Azure services.

Speaker Information

• Vamshi Kommineni
• Natalie Mao
• Saurabh Sensharma

Microsoft Events, Ignite 2025

Resources

• Migrate Application Workloads and Data to Azure
• Microsoft Ignite On-Demand Library

Chapters (from video)

• 0:00 – Overview of Azure Storage Families
• 2:03 – AI Workloads vs AI for Storage Structuring
• 3:02 – AI Training and Fine-Tuning Workflow
• 22:43 – Azure Container Storage & Kaito for Kubernetes AI
• 26:36 – Retrieval Augmented Generation and Data Preparation Steps
• 29:49 – Data and Network Security in Azure AI
• 32:09 – Demo: Azure Storage Center Management
• 36:24 – Demo: Azure Storage Integration with AI Services
• 37:20 – LangChain Integration on Azure

For further learning, browse the Microsoft Ignite session catalog.