Kostja
Zhang
“Three years of growth for other people’s AI products. Now building my own.”
From linguistics, to operator, to builder
Linguistics, two countries. Five years undergrad in Russia → five years study/work in Germany. A decade inside other people’s languages teaches the distance between “what the product means” and “the words a stranger would type.”
August 2022 — the operator seat. Joined a CV team doing ops + content. Midjourney arrived; ChatGPT followed. Sat inside Cutout.pro as it held a16z’s top-50 GenAI list. One hundred-plus AI/SaaS products across three years — each one sharpens the same lesson: distribution decides who survives.
July 2025 — the other hand. Picked up Vibe coding on the side. For the first time, prototype / landing page / growth loop fit in the same pair of hands. June 2026 — shipped Oginify in two days (first commercial product of my own). The consulting work continues in parallel; the building is what feeds it back.
“Two hands, one loop: operate to learn the market, build to keep the loop honest.”
From operating products to shipping them
Picked up Vibe coding in 2025 · June 2026 — shipped Oginify in two days. Oginify.com = free OG image generator, three a day, no signup. The growth playbook now writes itself with the product in the same pair of hands.
Visit Oginify ↗
Four takes I’ll defend
- 01
The operator who can also ship is unfair
Growth loop and product loop in the same hands compress feedback from weeks to hours. That window only really opened in 2025.
- 02
The growth function is undervalued
Algorithm > engineering > product > sales > ops on the pay chart. The market returns the same order on the revenue sheet.
- 03
Founders don’t lose to better models; they lose to silence
Most AI products are never seen by the user who would have loved them. Not a model problem — a distribution problem.
- 04
Going global is a re-learning problem, not a translation problem
Overseas channels are web-first and search-first. The playbook built around short video in China does not transplant — it has to be rebuilt from first principles.
The work, in numbers
“If you’re growing — or building — an AI product, I’d like to hear about it.”
More on what Alignify does → About
Your Product Deserves to Be Found.
Great AI products don't fail on quality — they fail on being invisible. We turn that around, one compounding system at a time.
Get started