praveenanil.dev

← Back to blog

Series · 5 chapters · ~16 min total

Build Your First AI App on Azure

Chapter 1: stand up a Foundry project, deploy a model, and call it from code with zero API keys — the foundation everything else in this series builds on.

Start with Part 1 →

Part 1 · 4 min readDeploy Your First Model on Microsoft FoundryChapter 1: stand up a Foundry project, deploy a model, and call it from code with zero API keys — the foundation everything else in this series builds on.

Part 2 · 3 min readGround the Model in Your Own Data — RAG with Azure AI SearchChapter 2: retrieval-augmented generation from scratch — index your documents in Azure AI Search, wire retrieval into the Responses API, and make the model answer from your facts with citations.

Part 3 · 3 min readGive the Model Hands — Tools and Function CallingChapter 3: function calling from first principles — let the model call your Python functions, mix in Foundry's hosted tools, and draw the safety lines before your app starts taking actions.

Part 4 · 3 min readStop Guessing — Evaluate Your AI App's OutputsChapter 4: build a golden dataset, score groundedness and relevance with AI-assisted evaluators in Foundry, wire evals into CI, and earn the right to change things without fear.

Part 5 · 3 min readShip It — Your AI App in Production on AzureChapter 5: containerize the app, deploy to Azure Container Apps with managed identity end-to-end, add content safety, rate limits, tracing, and cost controls — then go live.