How microsoftlearning/mslearn-ai-semantic-kernel works
Project type and purpose
• This repo is a hands-on tutorial consisting of five lab exercises (“Labfiles/01-05”) showing how to build generative AI apps using Azure OpenAI and the Semantic Kernel SDK.
• Each lab provides sample code in both C# (.NET) and Python, along with configuration files and data assets, so students can compare implementations side-by-side.
File structure
• Top-level
• .gitignore, LICENSE, README.md, Instructions/ (Markdown guides), index.md
• Labfiles/ (the actual lab projects)
• Per-lab folders (Labfiles/01-build-your-kernel, …, Labfiles/05-ai-assistant) each contain:
• A C-sharp/ subfolder: .csproj, Program.cs, appsettings.json, plugin or filter classes (FlightBookingPlugin.cs, Filters.cs), sample data (flights.json), build logs, etc.
• A Python/ subfolder: .env, kernel.py or prompts.py, plugin modules, data JSON, devops.py in the final lab.
• Configuration files (.env, appsettings.json) hold keys and endpoint URLs for Azure resources.
• Data files (flights.json) drive plugin examples.
• Instruction docs in Instructions/Labs/ map to the lab code under Labfiles/.
Key insights from the README
• The repo is educational, aligned with a Microsoft Learn path.
• Emphasizes use of Azure subscription, Azure OpenAI models, and Semantic Kernel as the orchestration layer.
• No single “running system” but five modular exercises demonstrating kernel-building, prompt execution, plugin creation, function filtering and an AI assistant pipeline.
Detail is key
• Break each lab into its core orchestration, plugin, and data components.
• Don’t just show “Backend” and “Frontend” – instead illustrate the kernel, function filters, prompt runners, and external AI service.