Microsoft has quietly built one of the better free on-ramps to generative AI here, and the fact that it lives on GitHub as an open-source course means it gets updated as the tooling moves rather than ageing the way a fixed video course would. The structure is its biggest strength, twenty-one lessons that start with what generative AI and large language models actually are, move through prompt engineering and prompt design, and then get you building, covering text generation apps, chat, retrieval over your own data, image generation, and on into agents and the practical concerns like security and improving your results. Each lesson is more than a video, you get a written walkthrough and code you can run, and the examples come in both Python and TypeScript, which is a thoughtful touch for web developers who do not want to context-switch into Python just to follow along. The thing to go in knowing is that it is a Microsoft course, so a good amount of the tooling leans on Azure OpenAI and the wider Azure ecosystem, and while they have made a real effort to include vanilla OpenAI and GitHub Models as alternatives so you are not forced into a cloud account, you will still spend time skimming past Azure-specific setup if that is not where you work.
It also assumes you can already write a bit of code, this is a course for developers getting into generative AI rather than an absolute-beginner introduction to programming, so non-coders should start somewhere gentler. And because it is fully self-guided there is nobody chasing you to finish, which for a 21-lesson course means you need a bit of discipline. Those caveats aside, for a free, maintained, genuinely hands-on tour of building with generative AI, this is one of the first things I would point a developer towards.