There is a moment most developers hit where they have played with ChatGPT in the browser and now want to actually call a model from their own code, and this is the course built precisely for that moment. DeepLearning.AI partnered with OpenAI to make it, Isa Fulford and Andrew Ng teach it, and the whole thing is structured around a Jupyter notebook that sits alongside the videos so you are not passively watching, you are changing prompts and rerunning cells and seeing how the output shifts in real time, which is by far the best way to learn this. It covers the practical principles that genuinely matter, things like writing clear and specific instructions, giving the model room to reason before it answers, and using the model for tasks like summarising, inferring, transforming, and expanding text, and then it walks through building a simple chatbot so you see the pieces fit together. The reason I like it as a starting point is that it cuts through the mystique around prompt engineering and reframes it as a normal, learnable development skill, and it does that in about an hour or two for free, which is an almost unbeatable ratio.
The honest framing is that it is exactly what it says, an introduction, so it will give you the foundations and the right instincts but not make you a specialist, and the examples are written against the OpenAI API specifically, so you will be translating a little if you work with another provider. You also need to be comfortable enough with Python to read and tweak the notebook, this is a developer course rather than a general-audience one. None of that takes away from the recommendation though, if you are a developer about to start building with large language models, do this first, and then go on to the longer courses about putting real applications together.