About this project
GenAI Academy is an independent, free course on generative AI. It covers 16 modules and 94 sections, starting from what artificial intelligence actually means and ending at mixture of experts and reasoning models.
What it covers
The curriculum runs across 5 units. The first builds the foundations: how machines learn, what a neural network is doing, why depth mattered. The second covers attention, transformers and large language models in detail. The remaining three move into applied work, retrieval-augmented generation, agents, fine-tuning, evaluation, deployment, and where the field is heading.
Nothing is locked. Everything is readable the moment you have an account, and an account exists only to remember where you got to.
How the lessons work
Every section is written twice. One version avoids notation entirely and leans on analogies; the other works through the actual mathematics. Both sit on the same page, so you can drop to the plainer one when you lose the thread and go back up when you have it.
Alongside that there are 41 diagrams built as interactive components rather than images, so you can change the inputs and watch what happens. Most sections include runnable code. Each one ends with a short quiz, and each module ends with a timed exam at 70% to pass.
There is also a tutor available on every section. It knows which section you have open, so you can ask for a different explanation without re-explaining your context. It is deliberately prevented from handing over quiz answers.
What isn't finished
All 94 sections are written and readable today. Several things around them are not done yet, and it seems better to list them than to let you discover them:
- •Section quizzes record your score but don't yet award XP.
- •The daily streak counter on the dashboard isn't counting yet.
- •Some badges listed on the achievements page can't be earned.
- •Writing quality is uneven. The earliest modules have been through more drafts than the most recent ones.
None of this affects the lessons themselves.
About the NVIDIA certification
The curriculum is structured around the published exam objectives for the NVIDIA-Certified Associate: Generative AI LLMs exam, because an external syllabus is a better guide than an invented one. Working through these modules covers that ground.
To be explicit about what this is not: this project has no affiliation with NVIDIA and no endorsement from them, and completing this course does not certify you in anything. The exam is sat with NVIDIA directly.
Cost
Free, with no payment page and no billing code anywhere in the project. The only running expense is the tutor, which calls a paid API. If that becomes a problem, the tutor gets a message limit before anything else gets a price. Should that ever change, existing accounts keep the access they have.