HomeBooks
Books
Each book starts from a problem you notice in the business, then explains the cause, the kind of fix, who needs it and what it takes. The technical chapters come last, for the engineers who build it.
Not sure which book fits? Describe your problem in a sentence.
13 chapters Kafka and event streaming From late reports and systems that disagree, to what event streaming is, who needs it, what it takes, and how to run Kafka well. 23 chapters AI agents in production How AI agents are built, deployed, monitored and rolled back once real users and real systems depend on them. 21 chapters LLMs in production: choosing a model, speed and cost How to choose a language model, why answers are slow, and what drives the cost of running one in production. 17 chapters Fine-tuning LLMs When fine-tuning a language model pays off, when retrieval or a better prompt is enough, and what fine-tuning takes. 14 chapters GPU clusters and AI infrastructure How GPU clusters for AI are sized, scheduled and kept busy, and what makes them expensive. 14 chapters Choosing a cloud and reading the bill How Google Cloud, AWS and Azure compare for real workloads, and how to read and reduce the bill. 13 chapters Analytics databases: ClickHouse, Postgres and warehouses Why dashboards slow down as data grows, how ClickHouse, Postgres and warehouses differ, and how to choose and migrate. 12 chapters Kubernetes cost and autoscaling Why a Kubernetes bill grows faster than usage, and how autoscaling and Karpenter bring it back in line. 8 chapters Docker in practice Running containers day to day: Docker on every platform, debugging containers that will not start, and cleaning up safely. 5 chapters RAG and retrieval How retrieval-augmented generation works, why AI assistants give wrong answers, and how to build retrieval that holds up.