Final Capstone
You've walked the whole arc: how LLMs work, the roadmap from zero to shipped, the lifecycle, the stack, evaluation, safety, fine-tuning, multimodal, the three scales of AI work, the decision frameworks, the production patterns, the career layer, eight shipped architectures, and (if you read it) the optional cutting-edge look ahead.
This capstone certifies it. 26 questions in the bank — each visit picks 12 at random, mixed across every chapter. The pass bar is higher than the chapter checkpoints: ≥ 75%. If you miss one, the result card links the page to revisit.
There's nothing here you haven't seen. If you passed every chapter checkpoint honestly, this is a victory lap.
Final capstone — whole-guide assessment
Pass to unlock the Next button belowPassed?
That's the whole guide. You have the mental models, the build path, the measurement discipline, the safety instincts, and the judgment layer. The only thing left is the thing no guide can do for you: ship something real, instrument it, and iterate.
- Build: start from Stage 0 if you haven't shipped the roadmap project yet.
- Look up: the Glossary is the reference backstop.
- Stay current: the Model snapshot carries the volatile facts.
Didn't pass?
The result card lists exactly which pages to revisit. Re-read them, retake (new questions each time), and don't sweat it — the bar here is deliberately higher than the chapter checkpoints.