AgentFlowAcademy

Track · For everyone

AI & LLM Foundations

What a model actually does, from neural networks to next-token prediction.

Free 11 lessons · 3 modules · 11 live now

Module 1 How models learn

  1. 01 AI, machine learning and deep learning Three nested ideas, and where LLMs sit. Start →
  2. 02 How a neural network learns Watch weights nudge toward the right answer. Start →
  3. 03 Training an LLM Pre-training, fine-tuning and learning from human feedback. Start →

Module 2 Inside an LLM

  1. 04 What is an LLM? Tokens in, one predicted token out, repeated. With a live temperature lab. Start →
  2. 05 Tokens and tokenizers Why “Chennai” is two tokens and what that costs. Start →
  3. 06 Embeddings Words as points in space; meaning as distance. Start →
  4. 07 Attention and transformers How every token looks at every other token. Start →
  5. 08 Sampling: temperature, top-k, top-p Turning probabilities into a choice. Start →

Module 3 Using LLMs well

  1. 09 Context windows and prompts What fits, what falls out, what the model really sees. Start →
  2. 10 Hallucinations and grounding Why models make things up, and how to stop it. Start →
  3. 11 Multimodal models Images, audio and video as tokens. Start →