The judgment layer between your business and its data.

I work where data problems are operational, not theoretical: revenue that has to be invoiced correctly, integrations that have to run at 3am, dashboards a board will make decisions from. The case files on this site are real engagements — sectors and numbers are accurate, names are withheld by design.

The pattern across them is the same. Understand the business mechanism first — a revenue-share ratchet, an attribution promise, a bidding loop — then build the smallest system that makes it reliable, observable, and boring. Warehouses that absorb new sources without redesign. Integrations that alert before customers notice. Numbers that agree with each other because they come from one place.

I'm trained as a physicist, which in practice means one habit: distrust of numbers until the model behind them is understood. It is a useful habit in data work.

Based in Tashkent, working across US, UK, and EU time zones — directly, or embedded with your team. Delivery in English, Russian, or Uzbek, including specifications and documentation your local team can actually use.