Blog

Notes from practice.

What I learn while leading an engineering team and building secure platforms, plus the games, communities and early web that shaped how I build. Short, concrete and written to be useful, not to be right.

All articles#

AI coding agents need a logging contract, not a better prompt

AI coding agents can pass the tests and still leave poor production evidence. A practical contract combines structured events, security rules and CI checks.

Read the article

Faster coding is not faster delivery

Notes on IBM Technology’s AI in the SDLC video: coding gains get absorbed by the delivery lifecycle, and the fix is redesigning the lifecycle around AI. With the video and one diagram.

Read the article

The future of work is Claude assisted by a human

We started with a human assisted by AI. The roles are quietly flipping, and the humans who thrive will be the ones who direct, verify and take responsibility.

Read the article

How I organize the Solutions engineering team

The operating model behind one team and one portfolio: named owners with backups, a single view of priorities and autonomy fed by context.

Read the article

Claude in the enterprise, from architecture to production

Why I keep coming back to Claude, and how an engineering team can use it across the whole lifecycle without giving up on security, confidentiality or judgment.

Read the article