AI & ML Academy

Content

Welcome to the AI and ML Academy (AIA)! This academy includes 6 modules, all of which are built on top of each other. It incorporates technical assets as well as best practices. We include common AI/ML scenarios that we have seen throughout various industries. The objective of the academy is to help you move faster and build effective solutions. The structure of the content is represented in the table below. Each session contains one or more presentations – with suggested labs and workshop content – designed to accommodate various skill levels.

AI & ML Overview
Azure OpenAI Generative AI, including GPT, Codex, DALL-E, and ChatGPT
Agentic AI Autonomous systems that plan, reason, and act with minimal human input, enabling dynamic, goal-driven task execution across business processes by integrating tools like Copilot, Semantic Kernel, and AutoGen
Scenario-based Services Document Intelligence, Azure AI Search, Metrics Advisor, Personalizer, Video Indexer, Anomaly Detector, and Power Virtual Agents
Customizable AI Azure AI Services, including Vision, Language, Speech, and customization of these Cognitive Services
Build Your Own ML Custom ML with Notebooks, Auto ML, Designer using Azure ML
ML Platform Train and Deploy models across a host of environments and compute types
ML Engineering in Production (MLOps) Azure DevOps, GitHub Actions, Kubeflow
 

Best Practices (newly added!)

Our latest best practices continue to learn more about Azure AI best practices:

 
 

Build 2025 AI Announcements and Updates

Summary of all announcements Book of News

To view all session recaps, please register for Build 2025.

Announcement Build Session Summary Session Number
Azure AI Foundry Agent Services The session on Azure AI Foundry Agent Service provided an overview of its capabilities, detailed how agents can be integrated using protocols like MCP and A2A, highlighted the use of Semantic Kernel and Process Framework for managing workflows, including integration with Microsoft 365 and Copilot Studio BRK149
Azure AI Content Understanding Azure AI Content Understanding streamlines content processing tasks by automating complex workflows, delivering structured and reliable data tailored to specific goals, optimizing processing performance with advanced reasoning capabilities, enabling efficient data extraction, and save time BRK169
Agentic Retrieval in Azure AI Search The session introduced Agentic Retrieval in Azure AI Search and showcased Azure AI Foundry as a comprehensive, widely adopted platform for building and integrating advanced, efficient AI-powered agentic search and generative applications, with real-world demos and resources to help developers continuously innovate and expand their solutions BRK142
Voice Live API in Azure AI Speech The session presented new APIs and features in Azure AI Speech—including real-time transcription, advanced text-to-speech, video translation, and the all-in-one Voice Live API—demonstrating with real-world examples how these unified speech capabilities empower developers to easily build accessible, customizable, and multi-modal voice-enabled AI solutions for diverse industries and global audiences BRK144
Azure AI Foundry API & SDK The session highlighted how Azure AI Foundry and its SDKs offer developers an end-to-end, fully integrated platform with robust models, seamless tools, and real-time deployment capabilities—demonstrated through practical invoice processing use cases—making it easier to build, manage, and scale comprehensive AI solutions while providing resources for quick adoption and further exploration BRK154
Visual Studio Code Extension The session demonstrated how to build, manage, and migrate AI agents from local development in VS Code to secure, scalable deployments in Azure using Azure SDKs, VS Code extensions, and tools like Azure Container Apps and CLI—making it easier for developers to enhance and protect their AI applications with robust cloud resources and practical guidance BRK117
Azure AI Foundry Observability The session showcased how integrating Azure AI Foundry with Azure Monitor enables developers to proactively monitor, analyze, and optimize AI agents—with powerful dashboards, alerts, and actionable insights—streamlining troubleshooting, performance tuning, and ensuring reliable, high-quality AI applications BRK168
Security and Governance The session detailed Microsoft's approach to building trustworthy, secure, and user-centric AI systems—highlighted by real-world banking applications, advanced evaluation and security measures like Microsoft Defender, and the importance of continuous improvement and cross-industry collaboration to ensure reliable, compliant, and effective digital transformation with responsible AI BRK145
Semantic Kernel + AutoGen The session showcased how Azure AI Foundry empowers developers to build advanced multi-agent applications—demonstrated through real-world use cases like Asus's interactive assistant—by providing integrated tools, seamless agent collaboration, and accessible resources, all while inviting community feedback to drive ongoing innovation BRK148
Azure AI Foundry Models The session highlighted how Azure AI Foundry streamlines regulatory compliance—particularly in financial services—by enabling the development and deployment of AI models and agents that automate compliance checks, suggest compliant alternatives, and offer practical workflow efficiencies, with ongoing improvements planned for broader accessibility and effectiveness BRK174
Model Router The session introduced new AI Foundry features for scaling AI with Azure Open AI services, addressed common scaling and performance challenges, provided practical demos and case studies, and offered guidance on leveraging these tools to maximize efficiency, value, and robust application deployment BRK178
New Fine-Tuning Capabilities The session showcased Azure AI Foundry's advanced fine tuning and distillation tools—highlighting new product features, real-world customer case studies, and improved workflows—to help users boost model performance and productivity BRK150
Foundry Local Foundry Local, a cross-platform, fast, secure, and private offline AI runtime for running models and agents locally on Windows and Mac; it integrates with Azure AI Foundry, leverages Microsoft assets, and will enable seamless switching between local and cloud AI for better performance, privacy, and customizability BRK146
Azure OpenAI The session unveiled cutting-edge new features—including a soon-to-launch, unique capability—for Azure Open AI, emphasizing same-day access to models, enhanced deployment, strong data protection, and compliance, while customer success stories and future plans reaffirmed Azure's leadership and ongoing innovation in enterprise AI BRK173
Multimodal Models The session introduced Sora—a powerful, state-of-the-art video generation model now available on Azure Open AI—spotlighting seamless API integration, robust trust frameworks, an interactive video playground, and real-world use cases that simplify high-quality video creation for enterprises and creators, with ongoing support and enhancements promised BRK170
Reasoning Models The session introduced advanced agentic reasoning models in Azure AI Foundry—highlighting their sophisticated problem-solving and pattern identification capabilities, practical implementation guidance, and associated higher complexity and cost—while encouraging attendees to explore and apply these advanced models for real-world impact BRK171
Optimizing Gen AI Apps The event introduced new AI Foundry features in Azure Open AI, demonstrating how they boost efficiency and scalability for deploying AI models, tackled common challenges with practical solutions and case studies, and encouraged attendees to leverage these advances for impactful large-scale AI applications BRK178

The post-event Developer Actions for Microsoft Build are available from this repository. This includes repositories organized by Build session, as well as Microsoft Learn content featured in Collections and Plans for continued learning in your own environment.

 
 

Vignettes (recently updated!)

Our ongoing vignettes continue the series by looking at the latest AI & ML topics:

GenAIOps Playbook Part 2
GenAIOps Playbook Part 1
 
 

Events

 
 

Specialization Program Guides

Preparation Guidance

 
 

Contributions

We welcome contributors to this project. Please use the GitHub links near the upper right and consider submitting pull requests or filing issues as needed. Curious how to contribute? See our Contribution Cheat Sheet.