Story
I started in QA, watching releases fail. Not in theory — at 2am, on the phone, while a production incident burned. I got obsessed with one question: why does shipping software have to feel like defusing a bomb every time?
So I spent ten years answering it. QA engineer to QA lead to engineering manager. At Zonatelecom I cut production incidents 3.4×. At Stenn I took release regression from 4–7 days to 20 minutes and moved nine product teams off a central QA queue onto shipping independently. At a regulated crypto company I cut production incidents by more than half. The pattern was always the same: take a delivery system that's on fire, rebuild the processes and ownership, hand back a machine that ships without heroics.
Then I wanted to know if I could build the whole machine myself. So I designed an AI-driven development pipeline and used it to ship two products end-to-end — from first idea to app stores and first paying customers — without writing the code by hand. Not a side experiment. Proof that AI, applied across the whole delivery lifecycle, changes what one person can ship.
I'm not picking a side between "engineering leader" and "solo founder." I think the interesting work lives in the gap between them. That's what I write about here.