I sat the CCAR-F exam, passed it with 1000/1000, and then had to explain to several friends how I'd actually prepared. This is that answer, written down properly.
I'm a developer with more than fifteen years of experience. When friends studying for the same exam asked what I'd used, the honest answer was a pile of scattered notes, a few free YouTube courses, and a very long trail of conversations with Claude and Gemini. It worked, but it wasn't something I could hand to anyone else.
So I rebuilt the preparation as something you can work through start to finish. Every lesson is a narrated walkthrough with a diagram that moves as the explanation does, and the whole course is written directly against the official exam guide's task statements rather than against someone's summary of them.
Building it was also the point. It was a good excuse to see how far you can push Claude Opus, ChatGPT Astra, and a handful of open-weight models working together on one long piece of work — the writing, the diagrams, and the narration all came out of that.
Five modules at the official exam weights, seventy-odd narrated lessons, and a set of hands-on exercises that run locally. It's free, there's no sign-up, and nothing here is behind an account.
This is an independent project. It isn't affiliated with, endorsed by, or reviewed by Anthropic, and passing anything here guarantees nothing about the real exam. Where this course and the official exam guide disagree, the guide is right — tell me and I'll fix it.
Questions, corrections, or something that reads wrong? I'd genuinely like to hear it.