Advanced

How frontier AI actually works.

This free advanced AI course explains how today's frontier models work, without a paywall or a PhD. In 14 short lessons you'll cover transformers, how LLMs are trained and aligned, retrieval and agents, efficient fine-tuning, diffusion models and reinforcement learning.

Each lesson you finish lights a star.

  1. Attention is all you needThe transformer architecture and the attention mechanism that powers every modern LLM.8 min · 2 quick checks
  2. How LLMs are trainedPre-training, supervised fine-tuning, and learning from feedback: the three-stage recipe.7 min · 2 quick checks
  3. RAG & AI agentsGiving models fresh knowledge with retrieval, and the ability to act with tools.7 min · 2 quick checks
  4. Fine-tuning on a budget: LoRA & quantisationHow to customise a billion-parameter model on a single GPU.6 min · 2 quick checks
  5. Diffusion: art from noiseHow Stable Diffusion and friends turn static into pictures, one denoising step at a time.6 min · 2 quick checks
  6. Reinforcement learning, deeperPolicies, values and rewards: the maths of learning by doing.7 min · 2 quick checks
  7. Scaling laws, evals & safetyWhy bigger got better, how we measure frontier models, and the open problems in keeping AI safe.7 min · 2 quick checks
  8. Agents: loops, tools and memoryAn agent is a language model in a loop that can act. How the loop works, how tools plug in, and why agents fail in surprising ways.8 min · 2 quick checks
  9. Thinking before answering: reasoning modelsWhy letting a model write out its thinking makes it smarter, and how reinforcement learning taught models to think for longer.7 min · 2 quick checks
  10. Keeping AI on our side: alignment and interpretabilityHow developers try to make models helpful, honest and harmless, how they test for danger, and how researchers are learning to read a model's mind.8 min · 2 quick checks
  11. How models are served: prefill, decode and the KV cacheNewWhat happens on the computer when you press send. Why replies stream out word by word, why output tokens cost more, and the tricks that make it all faster.8 min · 2 quick checks
  12. Mixture of experts: big models, small billsNewHow some of the largest models only use a small slice of themselves for each token, and what that saves and costs.8 min · 2 quick checks
  13. Distillation and the rise of small modelsNewHow a big model can teach a small one, why small models keep getting better, and when a little model is the right choice.7 min · 2 quick checks
  14. AI with a body: robots that learnNewWhy folding laundry is harder than passing exams, and how robots are starting to learn from simulation, demonstrations and language models.8 min · 2 quick checks
Read it offline

Save all of the Advanced level on this device for flights, trains and bad Wi-Fi.

Questions about the advanced level

Is there a free course on large language models?

Yes. The Advanced level here covers LLMs from transformers to training, RAG, agents and fine-tuning, free and with no sign-up. For hands-on code, pair it with the free Hugging Face LLM Course or Andrej Karpathy's Zero to Hero.

Do I need to code for the advanced level?

No. The lessons explain ideas in plain English with diagrams. Some include short optional Python examples that run in your browser.

What should I learn before the advanced level?

The Intermediate level, or similar background: gradient descent, neural networks, embeddings and how models are evaluated.