Most engineering starts on paper. A sketch, a rough diagram, a few lines of arithmetic on the back of a napkin, an argument you’re trying to talk yourself into or out of. By the time an idea becomes code, the interesting decisions have usually already been made — and lost. This blog is an attempt to keep some of that thinking around.

I’ll be writing about three things, mostly.

The first is systems design: how real software is structured, why it ends up looking the way it does, and the trade-offs that never make it into the README. The second is developer tooling — the small pieces of leverage that quietly change how a team works. The third is applied AI: not the breathless version, but the practical question of what machine learning is actually good for, where it’s being oversold, and what’s genuinely worth building right now.

My bias is toward the concrete. I’d rather walk through one real decision in detail than survey ten in the abstract. When I get something wrong — and I will — I’d like the writing to be specific enough that you can tell me exactly where.

If any of that sounds useful, you can subscribe via RSS or find me on GitHub. More soon.