// the take
A personal collection of numbered notebooks and scripts walking through Azure OpenAI, RAG, Semantic Kernel, LangChain, and APIM-fronted LLM setups. It's built for developers learning the Azure AI stack step by step, not for pulling into a real project.
The numbered folder structure (010, 020, 030, 100...) gives a genuine progression from raw SDK calls to RAG to agent patterns, which is more useful as a learning path than most scattered tutorial repos. It covers both Python and C# for the same concepts side by side, and the RAG folder ships with a real corpus of scraped Azure AI docs plus pre-chunked JSON, so you can run the retrieval examples without hunting down your own dataset.
Compiled build output (bin/obj, .dll, .pdb, .exe) is checked into git for at least one of the C# projects, which bloats the repo and will conflict the moment someone builds locally. A .env file is committed in 100_rag_intro, which is a real risk if it ever held live keys and sets a bad example regardless. There's no top-level README tying the folders together or explaining prerequisites, and the Infracost/Terraform manifest files suggest infra tooling that isn't documented anywhere in the tree, so you're left guessing how the pieces are meant to connect.
We read the README, the directory tree and a key source file in October 2026. Nobody pays to be listed. How we judge →