This chapter introduces key ideas behind Large Language Models (LLMs) and their growing role in AI-powered recommendation systems. It reviews the development of natural language processing (NLP), from early rule-based methods to modern transformer architectures like BERT and GPT, providing foundation for understanding how LLMs process unstructured text. Core concepts are covered, such as tokenization, embeddings, attention mechanisms, and retrieval techniques. The chapter also explains how LLMs are built and trained, including pre-training, fine-tuning, reinforcement learning with human feedback (RLHF). To connect theory with practice, the chapter provides hands-on tutorials that guide readers through tasks like tokenization, attention visualization, and retrieval pipelines—empowering both new learners and experienced users to apply LLMs to real-world challenges.