LLMs from Scratch - Practical Engineering from Base Model to PPO RLHF

Master the complete engineering process of building large language models from the ground up in this comprehensive 6-hour course using pure PyTorch. Begin with foundational transformer architecture concepts and progress through training a basic LLM, then advance to modern architectural improvements and scaling techniques. Explore cutting-edge approaches including Mixture-of-Experts (MoE) implementations, supervised fine-tuning methods, reward modeling frameworks, and reinforcement learning from human feedback using Proximal Policy Optimization (PPO). Gain hands-on experience with each stage of the LLM development lifecycle, from core transformer components through advanced alignment techniques, enabling you to build and customize your own language models with deep technical understanding of the underlying engineering principles.