WATCH. TRY. EXPLAIN.
Azure AI, in a few minutes.
Four short lessons to connect Microsoft Learn to your RAG project.
AT 0:00 · Start with a business question
Concept references · original explanation based on these sources.
- Approved sources and permitted readerslearn.microsoft.com · Prepare data for retrieval
- Azure Data and AI resource maplearn.microsoft.com · Azure Data and AI on Microsoft Learn
- Evidence-to-answer flowlearn.microsoft.com · Retrieve information and generate a response
The refund-policy question is a fictional teaching scenario.
LESSON 01 / 04
From source data to a useful AI answer
Give data, retrieval, and model services clear jobs, then test the source before improving the prompt.
Your place saves while watching and when paused. Reaching the end is tracked separately from “Reviewed”, your own checkpoint. Neither proves mastery.
TAKE IT INTO YOUR TERMINAL
Prepare documents you can trust
Validate a source inventory and build chunks that retain their provenance.
30–45 minutes · Notebook + Python scripts · No API key for the core exercise
A QUICK RECALL CHECK
The assistant quotes an expired policy accurately. What should you inspect first?
Practice mode: sign in before checking an answer to save your result.
Attempts made before assessment tracking was added were not stored.
FROM IDEA TO RUNNING CODE
Your build path.
Choose an agent project from Microsoft or OpenAI, or continue your foundations.
Download a starter. Make the tests pass. Measure what changes. Ship a small demo.
Time estimates are approximate. Core exercises use the Python standard library. Optional model experiments may need downloads, API credits, or GPU access.
Signed-in profiles and learning activity are visible to the app owner. Project notes stay private. Your profile & data