The AI-Native Engineer

A colleague gets promoted to Staff ahead of you. That night you open his files, not to be gracious about it, and what you find is not a method. It is early mistakes, the same question asked four ways, review comments, and a plan that kept changing while the work was still alive.

That is where this book starts.

It follows a payments engineer through eighteen months of working out which parts of her job survived, and it does not let her arrive at the answer cleanly. She is right in one chapter and wrong in the next, for reasons that look identical from the inside. She uses the right vocabulary to make the wrong call. Her mentor is wrong once, expensively, and stays in the book afterward.

Why a novel

Because the useful part of this shift is not a list of practices. It is judgment under time pressure, and judgment does not transfer as a list. It transfers by watching someone make the call and pay for it.

Books of advice about AI and engineering careers are not scarce. What is scarce is a cost attached to the advice. Every framework in this book arrives after someone has been wrong in a way that hurt, which is the only form in which any of it was ever learned.

What is inside

Ten chapters, roughly 25,000 words. It reads in an evening.

Then Field Notes: The Frameworks, the back matter: all ten frameworks organised for retrieval rather than for reading, one page each, with a practice attached. For the reader who finishes the story and comes back later with a problem in hand.

Who it is for

Engineers in the middle. Five to fifteen years in, senior enough to own something that matters, close enough to the work to notice that the ground moved and nobody said so. Managers and staff engineers read it for the chapters on teaching and on what cannot be verified from outside.

It assumes you can read code and does not require you to.

Get it

The framework pages on this site cover the ideas. The book is where they cost something.