What you'll understand after this part: How a large language model actually works — from tokens to transformers to attention mechanisms. How RAG retrieves knowledge. How agents reason and act. How embeddings capture meaning. How evaluations tell you if your AI system is working. Not at a hand-wavy level — at the level where you can make real engineering and product decisions.

What you'll build: An AI-powered data dashboard that lets you upload a CSV, ask questions in natural language, and get charts and insights. LLM integration, prompt engineering, and basic RAG — all integrated.

The teachers behind this part: Jay Alammar's visual transformer explanations, Andrej Karpathy's from-scratch intuition, Chip Huyen's production lens, Lilian Weng's research depth, Eugene Yan's system patterns, Hamel Husain's evaluation frameworks. We've synthesized the best explanations from the best minds and added the ARIA treatment so every concept is accessible without dumbing it down.