Generative AI is increasingly creeping in nearly any task I am working at these days, so it’s time to learn more about it. I am actually not interested in detail, how an LLM is internally working but how it can be applied and managed on a structural level. This means coding since we are not there yet that the model itself is tutoring beginners to grasp core concepts.

I am a fan of building things, and I can do this best in a new area when I get some guidance. The free online course LangChain for LLM Application Development does this perfectly. It comes with a video tutorial side-to-side with a jupyter notebook where you can follow along and make your own code experiments.

So, I listen, try myself, and … draw. Drawing is my reflection moment. Extracting key concepts from a lesson, visualize them in a consistent way, and keep the drawing somewhat consistent in style and depth over multiple lessons is for me the best way of recapitalizing and finding out whether I really got the idea.

A good side effect is that I have a visual reminder of the fundamentals which I can use later as my personal reference. Handy, isn’t it?