Chapter C · 3 of 4

Brain lab

LLM · tokens · RAG · agents — as toys

1. 01What is an LLM?7m
2. 02Tokens — how AI reads text6m
3. 2.5Training vs inference8m
4. 2.7Transformers in plain English9m
5. 2.9Serving Large Language Models9m
6. 03Context window suitcase6m
7. 3.1Memory & conversation state8m
8. 3.55Prompt caching8m
9. 04Temperature — safe vs creative6m
10. 05Embeddings — meaning as numbers7m
11. 06Vectors & similarity search7m
12. 07Overfitting playground6m
13. 7.5Multimodal AI8m
14. 7.6Diffusion models in plain English8m
15. 08RAG — look it up, then answer8m
16. 09Inside RAG — the pipeline8m
17. 9.2Chunking for RAG quality8m
18. 9.3Vector databases explained9m
19. 9.4Hybrid search for RAG9m
20. 9.5Fine-tuning vs RAG9m
21. 10Tools — when AI takes action7m
22. 10.1Structured outputs & JSON mode8m
23. 11Agents — think, act, repeat8m
24. 11.5Multi-Agent Systems9m
25. 12MCP — a standard plug for AI tools8m
26. 13Alignment and RLHF9m
27. 13.5Human in the loop8m
28. 14Evals and benchmarks8m
29. 14.5Reasoning models8m
30. 14.6AI Monitoring9m
31. 15Prompt injection & AI security9m
32. 16Choosing a model9m
33. 16.2Local LLMs & Ollama8m
34. 16.5Production AI Architecture10m
35. 16.6Computer-use agents10m
36. 16.7Text to SQL10m
37. 16.8Synthetic data10m
38. 16.9Knowledge graphs and RAG10m