Generative AI

Generative AI encompasses models that can create new content, including text, images, code, and more. At the core of modern generative AI are Large Language Models (LLMs) built on the Transformer architecture.1


Topics in This Section

TopicDescription
Attention MechanismsSelf-attention, cross-attention, multi-head attention
TransformersThe architecture behind GPT, BERT, and modern LLMs
Prompt EngineeringTechniques for effective LLM interactions
RAG SystemsRetrieval-Augmented Generation for grounded responses
Chunk EngineeringOptimal text segmentation for retrieval
Fine-TuningAdapting models to specific tasks (LoRA, QLoRA)
LLM EvaluationMeasuring quality, hallucination, and safety

Learning Path

Attention Mechanisms → Transformers → Prompt Engineering → RAG Systems → Fine-Tuning

  • Need NLP foundations? See 04 - NLP
  • Interested in RLHF? See 06 - Reinforcement Learning
  • Deploying LLMs? See 07 - MLOps & DevOps

References


  1. Vaswani, A., et al. (2017). Attention Is All You Need. Advances in Neural Information Processing Systems. https://arxiv.org/abs/1706.03762 ↩︎