— draft 37 —
About  ·  the long answer

A short bio, written reluctantly, in the third person.1

Ahmad Afzal is an AI product leader who has spent the last decade in that narrow, unglamorous strip between what research can do and what people will actually trust.

He writes about governance, accountability, and the quiet decisions — made on Tuesdays, usually — that decide whether a system behaves or just performs.

§  The longer thing

He's built products used by several million people, most of whom will never know his name, which is — he'd like you to believe — how he prefers it. He's written specs that became things, PRDs that became nothing, and at least one architecture diagram that his team still quotes when something is about to go catastrophically wrong.

His current obsession is evaluation: the unglamorous work of deciding, ahead of time, what "good" looks like — so that when the model does something strange at 2am, there is a grown-up in the room and not a screenshot on Twitter.

He is a warm presence in meetings, a slightly colder one in Google Docs, and genuinely enjoys a well-argued disagree-and-commit. He reads too much Donella Meadows. He has opinions about feedback loops. He knows this is the part of the bio that got too earnest, and he is going to leave it in anyway.

Off the clock he is learning — with mixed results — to sail small boats, to cook things that don't require a recipe, and to say "I don't know" more often in rooms where it would be strategic to say something else.2

§  Timeline

A selective chronology.

2024 — now
AI product leadership — reliability, evals, trust.
Building the tooling that lets teams ship AI features without the founder doing incident review at 2am. Mostly.
2021 — 2024
Head of Product, a company you've heard of once or twice.
Shipped to millions. Wrote fewer slide decks than was politically wise. Discovered the org chart is the product.
2017 — 2021
Product at a small studio; some consulting.
Learned that "it depends" is a complete sentence, and sometimes the best one in the meeting.
2014 — 2017
Early engineer & PM hybrid, pre-title inflation.
Worked on things that broke in interesting ways. Kept a notebook, which is most of how he still thinks.
before that
Studied systems, philosophy, and what rooms actually want.
Formal credentials exist and are boring. Ask him in person if it matters.
He is, on balance, more interested in
what a system rewards than what it says.
§  Miscellany

Things that are true, more or less.

Writes in

Obsidian, then a plain .md file, then nothing for three weeks, then everything at once.

Takes meetings

On walks when possible. Standing when not. Never at 4pm.

Reads

Systems thinking, old essays, the occasional novel he pretends to have finished.

Doesn't do

Cold DMs, quote-tweets, and "synergies." Has feelings about "pivot."

Keeps

A small garden of ideas. About 20% survive the winter.

Is not

A thought leader. Available to be one, however, in a pinch.3

§  Reading

On or near the desk, currently.

Open, being read

  • Thinking in Systems— Donella Meadows
  • The Systems Bible— John Gall
  • Selected essays on trust & institutions— various

Closed, pretending to have finished

  • Seeing Like a State— James C. Scott
  • The Design of Everyday Things— Don Norman
  • One novel he is lying about— name withheld
· · ·

Enough about him.

If any of the above sounded usefully adjacent to a problem you're working on — advisory, speaking, a careful second opinion — say hello. Warm intros preferred; cold ones read.