A comprehensive practitioner's reference for building autonomous AI systems covers the full stack from first principles to production deployment.
Its central thesis: building great agentic systems requires understanding every layer of the pipeline, not just one — LLM substrate, alignment and reasoning, then agentic training, RAG, memory systems, and context engineering.
Foundations include transformer architecture, GPU systems, training and fine-tuning (SFT, LoRA, MoE), model compression, and inference optimization.