4. Large Language Models

Generation, Reinforcement, Pretraining, and Engineering

1. Generative Models: Taxonomy & History
2. Autoencoders & VAE
3. GANs: Generator vs Discriminator
4. Conditional GANs & Pix2Pix
5. StyleGAN
6. Diffusion Models — DDPM from Scratch
7. Latent Diffusion & Stable Diffusion
8. ControlNet, LoRA & Conditioning
9. Inpainting, Outpainting & Editing
10. Video Generation
11. Audio Generation
12. 3D Generation
13. Flow Matching & Rectified Flows
14. Evaluation: FID, CLIP Score
15. Visual Autoregressive Modeling (VAR): Next-Scale Prediction
16. MDPs, States, Actions & Rewards
17. Dynamic Programming
18. Monte Carlo Methods
19. Q-Learning, SARSA
20. Deep Q-Networks (DQN)
21. Policy Gradients — REINFORCE
22. Actor-Critic — A2C, A3C
23. PPO
24. Reward Modeling & RLHF
25. Multi-Agent RL
26. Sim-to-Real Transfer
27. RL for Games
28. Tokenizers: BPE, WordPiece, SentencePiece
29. Building a Tokenizer from Scratch
30. Data Pipelines for Pre-Training
31. Pre-Training a Mini GPT (124M)
32. Distributed Training, FSDP, DeepSpeed
33. Instruction Tuning — SFT
34. RLHF — Reward Model + PPO
35. DPO — Direct Preference Optimization
36. Constitutional AI & Self-Improvement
37. Evaluation — Benchmarks, Evals
38. Quantization: INT8, GPTQ, AWQ, GGUF
39. Inference Optimization
40. Building a Complete LLM Pipeline
41. Open Models: Architecture Walkthroughs
42. Speculative Decoding and EAGLE-3
43. Differential Attention (V2)
44. Native Sparse Attention (DeepSeek NSA)
45. Multi-Token Prediction (MTP)
46. DualPipe Parallelism
47. DeepSeek-V3 Architecture Walkthrough
48. Jamba — Hybrid SSM-Transformer
49. Async and Hogwild! Inference
50. Speculative Decoding and EAGLE
51. Gradient Checkpointing and Activation Recomputation
52. Prompt Engineering: Techniques & Patterns
53. Few-Shot, CoT, Tree-of-Thought
54. Structured Outputs
55. Embeddings & Vector Representations
56. Context Engineering
57. RAG: Retrieval-Augmented Generation
58. Advanced RAG: Chunking, Reranking
59. Fine-Tuning with LoRA & QLoRA
60. Function Calling & Tool Use
61. Evaluation & Testing
62. Caching, Rate Limiting & Cost
63. Guardrails & Safety
64. Building a Production LLM App
65. Model Context Protocol (MCP)
66. Prompt Caching & Context Caching
67. Agent State Machines — Graphs, Nodes, Checkpoints
68. Agent Framework Tradeoffs