An engineer’s journal on AI models and systemsResearch / Engineering / Practice

Adam’s AI engineering notebook

AI, measured in the wild.

I’m Adam - a software engineer for ten years, hooked on machine learning. I write up my experiments with the method left in: benchmarks, tutorials, and field notes for anyone curious how the models behave in practice.

Currently on the bench

What I’m testing right now.

Updated Jul 15, 2026

Nothing on the bench yet - the first experiments are warming up.

How I write here

Three rules I write by.

01

First-hand or not at all

Everything here comes from something I ran, built, or broke myself - not from paraphrasing announcements.

02

Show the work

The setup, the reasoning, and the dead ends stay in the post - whether it’s a benchmark or a tutorial.

03

Leave something usable

A number to cite, a script to run, or a mental model to keep - every post should hand you at least one.