.NET AI with Semantic Kernel

What You'll Learn

• Set up and configure Semantic Kernel with Azure OpenAI
• Build reusable plugins with KernelFunction
• Implement vector store memory for long-term context
• Orchestrate multi-step AI workflows
• Apply production patterns: error handling, retries, and telemetry

Prerequisites

• C# and .NET 8+
• Basic understanding of AI concepts (see AI Foundations path)
• Azure OpenAI or OpenAI API access

Learning Path Articles

1. 1

   What Is Semantic Kernel? Beginner Guide for .NET Developers

   New to Semantic Kernel? Learn what it is, where it fits in the Microsoft AI stack, and how .NET developers use plugins, memory, and function calling in C#.

   Beginner 12 min read university
2. 2

   Semantic Kernel .NET Setup: NuGet to Streaming Chat in 30 Minutes

   Your first Semantic Kernel app: NuGet install, Azure OpenAI config, streaming chat, plugins, and function calling. Working .NET 10 project in 30 minutes.

   Beginner 14 min read workshop
3. 3

   Build Semantic Kernel Plugins in C#: [KernelFunction] + DI + Auto Calling

   Step-by-step: create SK plugins with [KernelFunction], wire up DI, and let the LLM auto-invoke your C# tools. Plugins work in SK and Agent Framework.

   Intermediate 14 min read university
4. 4

   Semantic Kernel Memory + Vector Stores: Azure AI Search & Cosmos DB in C#

   Connect Semantic Kernel to Azure AI Search and Cosmos DB for RAG. Three memory levels, embedding setup, and vector store connectors in C#.

   Intermediate 11 min read university
5. 5

   Semantic Kernel Filters: Add Caching, Logging & Safety to AI Calls in C#

   Implement IFunctionInvocationFilter, IPromptRenderFilter, and IAutoFunctionInvocationFilter in C#. Cache AI responses, log token usage, detect prompt injection.

   Intermediate 14 min read university
6. 6

   Migrate Semantic Kernel Agents → Microsoft Agent Framework (.NET)

   Side-by-side migration guide: Semantic Kernel Agents to Microsoft Agent Framework in .NET. What transfers directly, what changes, and when to migrate.

   Advanced 14 min read university
7. 7

   SK Planners Are Deprecated — Migrate to Auto Function Calling (C# Code)

   SK planners are deprecated. Migrate to FunctionChoiceBehavior.Auto() with MaxAutoInvokeAttempts control, IAutoFunctionInvocationFilter, and circuit-breaking. C# before/after code.

   Intermediate 13 min read university
8. 8

   Manage Semantic Kernel ChatHistory in C#: Sliding Window, Summarization & Cosmos DB

   Stop ChatHistory from blowing up your token budget. Sliding window, hybrid summarization, token-aware truncation, and Cosmos DB persistence in C#.

   Intermediate 14 min read university
9. 9

   Semantic Kernel Streaming in C#: Token Tracking, Content Filters & Blazor

   InvokeStreamingAsync vs GetStreamingChatMessageContentsAsync, token usage workaround, content filter detection, and Blazor Server patterns. Working C# code.

   Intermediate 13 min read university