Large language models are the application layer of transformer-based AI. These lessons cover how LLMs are trained, aligned to human intent, and deployed in production systems — from the raw next-token prediction objective through the agent architectures that extend LLM capabilities beyond text generation.