Getting started with generative AI in Azure

With the conceptual groundwork in place, you now move into the platform where your AI work will happen. This module introduces Microsoft Foundry — the current name for what was previously Azure AI Foundry — and takes you through setting up your own environment, navigating its key interfaces, and configuring AI models for practical use. The core of the module is experimentation. You learn what AI experimentation means and why a structured approach to it matters, then run your own experiments: adjusting generation parameters such as temperature, top-p, and reasoning effort, and observing directly how each setting shapes model output. The hands-on labs are scaffolded so that each step builds on the last. By the end, you will be confident navigating Foundry, able to set up and run experiments independently, and equipped to interpret results — a foundation you will build on directly when you begin developing your first AI application.

This is where you build. The module takes you from design to a working text generation application, with each lesson scaffolded so that you understand the choices you are making before you make them. The first lesson focuses on the blueprint: you examine what makes a well-designed text generation application, choose an appropriate model for your use case, and understand how the API layer will connect your backend to the Foundry endpoint. You then get hands-on in a re-recorded screencast on the current Microsoft Foundry portal, and build your application's core framework in a structured practice activity — with a pre-built UI provided so your attention stays on the AI integration. The second lesson shifts to testing and refinement. You learn how to evaluate your application's outputs, apply prompt engineering techniques to improve them, and use the practice quiz and graded assessment to confirm your understanding before moving ahead. Note: this module deliberately does not cover fine-tuning or MLOps — those are intermediate-to-advanced topics outside the course scope. The focus is on building a functional, well-tested application using pre-built models.

Building a capable AI application is only part of the job — building one you can stand behind is the other. This final module weaves responsible AI directly into your development practice rather than treating it as an afterthought. First Lesson gives you the ethical grounding: what the core principles of responsible AI are, how bias and hallucinations arise and why they matter, and how to create a practical evaluation framework. The lesson sequence has been reordered so that the ethical framework comes before the bias-and-hallucinations deep dive, giving you the principles before the problems. A re-recorded screencast on the current Microsoft Foundry portal then shows you Azure's built-in responsible AI toolkit in action. Second Lesson moves into application: you implement content filters and prompt injection defences in a hands-on activity, work through a coach dialogue that connects principle to practice, and complete the final project — an enhanced, documented text generation application that integrates both technical functionality and the ethical safeguards you have just learned. A role play and graded assessment close the module, giving you a chance to demonstrate readiness before the course summary.