AI-500 Labs
Microsoft Learn Sandboxes require you to bring in your own subscription.
This repo is under active development alongside the beta exam; new labs may be added over time.
Design stateful agentic loopsDesign and test stateful agentic loops for the Contoso Capital Investment Research Platform.
Implement advanced multi-agent orchestrationImplement and test advanced multi-agent orchestration for the Contoso Capital Investment Research Platform.
Apply task decomposition and collaborationApply and test task decomposition and agent collaboration for the Contoso Capital Investment Research Platform.
Design enterprise agent-to-agent communicationDesign and test enterprise agent-to-agent communication for the Contoso Capital Investment Research Platform.
Design advanced prompting strategiesDesign and test advanced prompting strategies for the Northwind Health Clinical Intelligence System.
Build enterprise tool ecosystems with MCPBuild and test an enterprise Model Context Protocol tool ecosystem for the Northwind Health Clinical Intelligence System.
Implement advanced RAG with Azure AI SearchImplement and test advanced retrieval-augmented generation with Azure AI Search for the Northwind Health Clinical Intelligence System.
Design multi-agent memory architecturesDesign and test a multi-agent memory architecture for the Northwind Health Clinical Intelligence System.
Implement CI/CD for multi-agent systemsImplement and test CI/CD controls for the Fabrikam Code Review and Deployment Agent System.
Secure multi-agent systems with zero trustDesign and test zero-trust security controls for the Fabrikam Code Review and Deployment Agent System.
Scale responsible AI governanceDesign and test responsible AI governance controls for the Fabrikam Code Review and Deployment Agent System.
Govern the enterprise agent lifecycleDesign and test enterprise agent lifecycle governance for the Fabrikam Code Review and Deployment Agent System.
Implement distributed observabilityImplement and test distributed observability for the Adventure Works Customer Intelligence Platform.
Design multi-agent evaluation frameworksDesign and test a multi-agent evaluation framework for the Adventure Works Customer Intelligence Platform.
Optimize multi-agent performance and costDesign and test performance and cost controls for the Adventure Works Customer Intelligence Platform.
Design human-in-the-loop approval workflowsDesign and test human-in-the-loop approval workflows for the Adventure Works Customer Intelligence Platform.
Debug multi-agent production incidentsDiagnose and test multi-agent production incident response for the Adventure Works Customer Intelligence Platform.