LLM Engineering: Transformers, Agents & Production Apps

The whole LLM pipeline, end to end — from the transformer up to a deployed app. LLM Engineering: Transformers, Agents & Production Apps is a comprehensive engineering reference for building with large language models the way practitioners actually do: architecture first, then the training and alignment stack, then agents, then production.

You'll start inside the transformer — attention, tokenisation, and embeddings — then move through pre-training at scale, fine-tuning, and alignment (RLHF, DPO, and beyond). From there the book turns applied: prompt engineering, retrieval-augmented generation, LLM agents and multi-agent systems, long context and memory, inference optimisation and serving, and the APIs and orchestration frameworks that tie it all together. The final chapters cover evaluation and benchmarking, safety, guardrails and red-teaming, and production architecture and LLM MLOps — everything between a model and a reliable application.

You will learn to: - Understand transformer architecture, tokenisation, and embeddings deeply - Pre-train, fine-tune, and align models (RLHF, DPO) - Build RAG systems, agents, and multi-agent workflows - Optimise inference and serve models efficiently - Orchestrate LLM apps with modern frameworks and APIs - Evaluate, secure, and operate LLMs in production

Built for engineers who want the full picture — from the Math of attention to the MLOps of deployment.