Skip to content
SDB

AI Tutorials

Generative AI

Master generative AI from first principles — covering model families, transformers, LLM training, fine-tuning, prompt engineering, RAG, multimodal AI, evaluation, and guardrails with production-ready code.

13 chapters · 375 min

← All AI tutorials

From diffusion models to LLMs, RAG, and responsible deployment

  1. Ch. 01

    Different Generative Models

    A taxonomy of generative AI: GANs, VAEs, diffusion models, and autoregressive transformers

    beginner · 20 min

    Read →
  2. Ch. 02

    Transformer Architecture

    Self-attention, positional encoding, and the building blocks powering modern AI

    intermediate · 30 min

    Read →
  3. Ch. 03

    Large Language Models

    How LLMs are trained, what emerges at scale, and how to use them effectively

    intermediate · 25 min

    Read →
  4. Ch. 04

    Fine-Tuning & Adaptation

    LoRA, QLoRA, and parameter-efficient techniques for specializing LLMs

    intermediate · 25 min

    Read →
  5. Ch. 05

    Prompt Engineering

    Systematic techniques to guide LLMs toward better outputs

    beginner · 25 min

    Read →
  6. Ch. 06

    Retrieval-Augmented Generation (RAG)

    Grounding LLM responses in external knowledge with vector search and retrieval

    intermediate · 30 min

    Read →
  7. Ch. 07

    Multimodal Concepts

    Vision, audio, and unified models that understand and generate across modalities

    intermediate · 25 min

    Read →
  8. Ch. 08

    Responsible AI & Governance

    Bias, fairness, transparency, and the frameworks governing AI deployment

    beginner · 25 min

    Read →
  9. Ch. 09

    Key Evaluation Metrics

    Measuring LLM performance across reasoning, generation quality, and task-specific benchmarks

    intermediate · 25 min

    Read →
  10. Ch. 10

    Guardrails

    Building safety, quality, and compliance controls into production AI systems

    intermediate · 25 min

    Read →
  11. Ch. 11

    Diffusion Models Deep Dive

    DDPM math, DDIM sampling, classifier-free guidance, and LoRA for image generation

    advanced · 42 min

    Read →
  12. Ch. 12

    Multimodal AI

    CLIP, DALL-E 3, Gemini, GPT-4V, and image-text retrieval — the dominant direction of frontier AI

    intermediate · 38 min

    Read →
  13. Ch. 13

    LLM Alignment & RLHF

    PPO from human feedback, DPO, Constitutional AI, and reward modeling — why modern LLMs behave as they do

    advanced · 40 min

    Read →