How I use AI

Acceleration with judgment retained.

Hiring teams ask about AI, so here is the direct answer: I use it daily, it makes my delivery work faster, and the judgment, verification, and accountability stay with me.

01 / The short answer

Three working rules

02 / At work

A concrete example

On my current program I used Gemini to synthesize scattered test documentation, logs, specs, and field notes from a multi-day, continuously recording test window into a single event timeline for sponsor capability reporting.

The tool made real mistakes reading diagrams and device numbers. I caught them, corrected them, and verified the synthesized timeline against the source material before anyone relied on it. That is the judgment part of the work, and it is the part that does not get delegated.

03 / On my own delivery

Directing an AI-implemented build

My delivery portfolio is a working system I delivered by directing an AI-implemented build through a governed lifecycle: charter, RAID, RACI, MoSCoW, ADRs, and IEEE-829 verification and validation, with AI output held to the same review gates a technical program would enforce. This website was delivered the same way, run as a governed engagement with Jira, gate sign-offs, and an enforced quality gate in CI.

04 / Where the lines are

Boundaries