Packt Publishing · 2021 — 2026
From /etc/fstab to Model Context Protocol. A running log of the books I've technically reviewed for Packt — what they cover, and why reviewing them is the best way I know to learn a technology in depth.
I did my first credited technical review for Packt in 2021. It was a lot of fun — and it turns out to be one of the best ways I've found to learn a new technology in real depth. You don't get to skim. You have to run the code, break the examples, and argue with the author about whether that command actually works on RHEL. Plus, I got my name in print.
Since then I've been asked back for non-credited reviews, which are their own kind of fun: you get an early copy of the book months before the public does. Reviewing Operational AI with Docker meant reading about Docker Model Runner and agentic MCP integration before most of the industry had heard of it.
Most recently I was asked to do another credited review — this time on an upcoming Anthropic Claude title. That one I am genuinely excited about.
Newest first. Filter by how the review was credited.
The shelf starts at bare-metal Linux and ends at agentic AI. The through-line never changed: will this hold up in production at 3am?
Not copyediting. A technical reviewer is the person who has to answer, chapter by chapter:
| Check | What it means in practice |
|---|---|
| Does the code run? | Every listing, on a clean box, at the stated versions. Not "does it look right." |
| Is the command correct? | Distro drift is real — systemctl flags, package names, and default paths differ across RHEL, Debian, and SUSE. |
| Is the concept sound? | Does the mental model the author is teaching survive contact with production? |
| Is anything missing? | The step the author does automatically and forgot to write down. That's the one that costs a reader two hours. |
| Is it current? | A cloud console screenshot has a shelf life measured in months. |
Comments go back to the author and editor as inline annotations plus a per-chapter questionnaire and score. The author decides what to take. The best exchanges are the ones where you're both wrong at first.
20+ years of infrastructure and platform engineering — bare-metal Linux through multi-cloud Kubernetes — now pointed at agentic AI systems. Which is why the shelf reads the way it does.
Reviewing for Packt? If you're an author or editor looking for a technical reviewer on Linux, Kubernetes, cloud infrastructure, DevSecOps, or agentic AI systems — open an issue or reach me on LinkedIn.