Azure Storage for AI workloads​ | OD870

From the creator

Learn how Azure Storage powers AI inference at scale. This session explores how to securely bring enterprise data to AI models, accelerate AI workloads with high-performance storage, and reduce GPU idle time through faster model loading and optimized data access. See how Azure Storage integrates with Microsoft and open-source AI frameworks to improve performance, lower costs, and enable scalable agent-based applications. 𝗦𝗽𝗲𝗮𝗸𝗲𝗿𝘀: * Saurabh Sensharma * Vishnu Charan TJ * Saloni Sonpal 𝗦𝗲𝘀𝘀𝗶𝗼𝗻 𝗜𝗻𝗳𝗼𝗿𝗺𝗮𝘁𝗶𝗼𝗻: This is one of many sessions from the Microsoft Build 2026 event. View even more sessions on-demand and learn about Microsoft Build at https://build.microsoft.com OD870 | English (US) | Cloud platform & data Pre-recorded #MSBuild Chapters: 0:00 - Introduction to Azure Storage for AI workloads and session overview 00:01:20 - Framing the connection between Azure Storage and AI workloads 00:03:57 - Azure Storage integration with AI frameworks and infrastructure 00:05:20 - Understanding agentic inference and storage requirements 00:07:11 - Optimizing inference with prompt caching 00:10:10 - Three tiers of prompt caching in Microsoft Foundry 00:13:05 - Deep dive into Azure Context Cache design and enterprise benefits 00:17:47 - Fast model loading and distribution using Azure Blob and Streamer 00:25:05 - Bringing enterprise data into AI inference with integrated tools 00:29:00 - Connecting Azure Storage to Foundry IQ and Storage Center announcements

Published 2026-06-12 · Watch on YouTube ↗
Availability and playback restrictions are controlled by the original channel.