Prompt to Production: A Practitioner's Guide to DevOps and SRE in the AI Era - Softcover

Inhaltsangabe

The tools are real. So are the failure modes. This book is about the difference.

At three in the morning, mid-incident, an LLM handed me the right hypothesis in two minutes. A week earlier, the same model confidently invented a Terraform flag that does not exist. Both stories are true, and a useful book about AI in operations has to hold both at once.

I have spent sixteen years in infrastructure: university sysadmin, banking networks, telco-scale BGP, security, and now leading a DevOps team. This is the book I wanted to hand my own engineers: a practitioner's guide to LLMs in DevOps, SRE, and platform engineering, current as of mid-2026, with no vendor slides and no doom.

Inside:

• How LLMs actually work, in plain words, so you can predict where they fail
• The 2026 model landscape: frontier families, open-weight models, and local deployment for regulated environments
• Prompting fundamentals, and the few techniques that still matter
• Interaction modes: chat, IDE, terminal agents, PR-level agents, and when each is the right tool
• Context engineering, spec-driven development, skills and hooks
• Safety: what you may paste where, vendor terms, prompt injection, and the new agent attack surface
• The applied core: Infrastructure as Code, CI/CD and scripting, observability and incident response, security and compliance work, documentation
• Cost, measurement, and team adoption, with numbers instead of feelings
• The path to autonomy: agents, the verification law, and what stays human
• Two appendices: running local LLMs (hardware, models, tools, the team box for regulated environments) and copy-ready templates (a full AGENTS.md, a complete skill, a working terraform plan-gate hook, the one-page AI policy, the internal model card)

Two running case studies keep it concrete: a mid-size SaaS on AWS and a regulated bank running local models. Every chapter ends with a low-risk exercise to try on Monday and a short list of what to watch out for.

Who this is for: working DevOps engineers, SREs, platform engineers, sysadmins, and their team leads. If you can read a YAML file and have ever been paged, you are the reader. No ML background needed.

Who this is not for: anyone who wants AGI prophecy or a promise that AI will replace ops teams. It will not. It is changing what the job looks like, and this book is about ending up on the right side of that change.

Die Inhaltsangabe kann sich auf eine andere Ausgabe dieses Titels beziehen.