AI-500: Multi-Agent AI Solutions Expert

Articolo Completo

Autore:

8.6 (302 856)

15 nov 2026, 19:03

What you'll learn

⚡ Architecting Multi-Agent Solutions: Explain Foundry, GenAIOps, models, deployment options and pricing, and deploy your own Foundry project.
⚡ Prompts and Fine-Tuning Strategy: Manage prompt versions and variants, and plan a fine-tuning strategy for agents and models.
⚡ Tools and MCP: Build agents in the portal and in Python with web search, code interpreter, OpenAPI and MCP tools, and combine them.
⚡ Memory, Context and Knowledge: Design retrieval with chunking, embeddings and Azure AI Search, and ground agents with Foundry IQ.
⚡ Multi-Agent Orchestration: Orchestrate agents with Foundry workflows and Microsoft Agent Framework, including human-in-the-loop approval.
⚡ Monitoring and Observability: Route Foundry telemetry to Log Analytics, query it with KQL and trace agents.
⚡ Evaluation and Continuous Improvement: Design test datasets, score models and agents with built-in evaluators, and optimize an agent.
⚡ Securing Multi-Agent Solutions: Apply Zero Trust, identity, network controls and guardrails in Foundry, and protect AI with Defender for Cloud.
⚡ Guardrails and Responsible AI: Apply responsible AI principles and recognize AI risks such as prompt injection, data poisoning and overreliance.
⚡ Deploying Multi-Agent Solutions: Explain Bicep, safe rollout with blue/green deployments, and CI/CD pipelines with GitHub Actions.

Requirements

❗ Willingness to learn cool stuff!

Description

This course contains the use of artificial intelligence.
Multi-Agent AI Solutions Expert (AI-500) by Christopher Nett is a hands-on course that teaches you how to design, build, evaluate, secure and deploy multi-agent AI solutions on Microsoft Foundry, organized around the skills measured on the AI-500 exam.
You start with the Azure and generative AI basics, then learn the Foundry platform and deploy your own Foundry resource and project. From there you build agents with tools and knowledge, orchestrate them with workflows, and take them to production with monitoring, evaluation, security, guardrails and deployment practices.
By the end of this course, you will be able to architect and run multi-agent solutions on Azure and prepare for the AI-500 exam.
1. Basics - Azure covers regions and availability zones, the resource hierarchy, the core Azure services, and Entra ID with Azure RBAC. You set up your subscription with a budget alert and work with Azure through the portal, the Azure CLI, PowerShell, Cloud Shell and Copilot.
2. Basics - GenAI covers how large language models work, how to choose a model and its parameters, prompt engineering, embeddings, retrieval-augmented generation and grounding, and what AI agents and the Model Context Protocol are.
3. Architecting Multi-Agent Solutions covers Microsoft Foundry and GenAIOps, the Foundry resource and project architecture, Foundry Models, model deployment options and pricing. You create your own Foundry resource and project and deploy a model, which the later labs build on.
4. Prompts and Fine-Tuning Strategy covers prompt operations and version control, and when and how to fine-tune agents and models.
5. Tools and MCP covers Foundry Agent Service. You build agents in the portal and in Python, give them web search, code interpreter, OpenAPI and MCP tools, and combine the tools in one agent.
6. Memory, Context and Knowledge covers chunking, embeddings and retrieval strategies, Azure AI Search, and Foundry IQ knowledge bases for agents.
7. Multi-Agent Orchestration covers Foundry workflows, including human-in-the-loop approval.
8. Monitoring and Observability covers telemetry, tracing and cost. You create a Log Analytics workspace and configure Foundry telemetry.
9. Evaluation and Continuous Improvement covers evaluation strategy and test dataset design. You score a model with the built-in evaluators of the Azure AI Evaluation SDK and optimize an agent with the Agent Optimizer.
10. Securing Multi-Agent Solutions covers Zero Trust and identity, networking and content safety in Foundry, and Microsoft Defender for Cloud, which you configure to protect your AI services.
11. Guardrails and Responsible AI covers Microsoft's responsible AI principles and the common AI risks, including prompt injection, training data poisoning and overreliance.
12. Deploying Multi-Agent Solutions covers Bicep, safe rollout with blue/green deployments, and CI/CD with GitHub Actions.
Most labs build on the Foundry resource and project you create at the end of Architecting Multi-Agent Solutions, so complete that lab before the later sections. If you already know Azure and generative AI, go from the Lab Setup article straight to Architecting Multi-Agent Solutions. The course is built for experienced developers, AI engineers and architects who build agentic solutions on Azure and want to prepare for the AI-500 exam.
This course contains promotional materials.

Who this course is for

⭐ AI Engineers
⭐ AI Architects
⭐ Cloud Engineers
⭐ Software Engineers
⭐ Cloud Architects

Homepage

https://www.udemy.com/course/ai-500-multi-agent-ai-solutions-expert-s

Stai guardando "AI-500: Multi-Agent AI Solutions Expert" Su Libreria Oltre Il Paradiso

Stai guardando AI-500: Multi-Agent AI Solutions Expert!. Questo contenuto appartiene alla categori Corsi ed è stato aggiunto il Oggi, 19:03. Non dimenticare: Oltre il Paradiso libreria.

★ Offerta Premium FileStore

★ Come trovarci sempre

Commenti

Il commento deve contenere almeno 20 caratteri. Ti invitiamo a rispettare gli altri utenti.

Non ci sono ancora commenti. Lascia il primo commento!