Introduction to AI in Azure: Review

The second day, the instructor walked through a slide showing the Azure AI Vision service. Someone asked if we would build something with it. The answer was, “It’s covered in the lab, but we don’t have time to code it live.” That pretty much sums up the entire 24 hours. This isn’t a course; it’s a narrated PowerPoint deck you have to pay a partner a fortune to sit through.

I was expecting a guided tour of the Azure SDKs for a few core services—maybe train a tiny custom vision model, call the speech-to-text API from a notebook, something. Instead, we got a firehose of service names and marketing descriptions. Computer Vision, Custom Vision, Face, Form Recognizer, Video Indexer... We learned what they are, what the marketing page says they do, and which Azure portal screen to click on. The so-called 'labs' were pre-scripted, fill-in-the-blank exercises that didn't require understanding. You could pass them by just matching names from the slides.

As someone who's spent years writing backend code, this felt like a sales pitch aimed at a manager who needs to approve a budget. It's all about the 'what,' never the 'how.' For the price of this instructor-led session, I could have paid for a year of a competitor's full-stack AI program and actually shipped a project. It's purely for getting the AI-900 certificate on your company's dime. If you're paying for this yourself to learn how to build, stop. You're in the wrong place.