I build digital products: web services, Telegram bots and automation in TypeScript, sometimes in Python. Below are my projects, stack and contacts. Separately there is stuff: things I make for myself and out of curiosity.
Working on a VPN service and a payments platform.
updated · October 2026I write for Node.js in TypeScript, and Python less often. Mostly these are services that take something in, count it and pass it on: payments, events, subscriptions. I work on a product team, and I automate the routine around it — data collection, reports, notifications — with bots and scripts. I like solutions that ship fast and can be checked against numbers.
Product work in a team: payment flows, user onboarding, integrations. On the side I automate routine work: data collection, reports, bots.
I start with the simplest solution that works and add complexity only when there is no way around it. I judge the value by the numbers. A good part of my code is written with AI agents: spec first, then implementation, then tests. Complex tasks get a cross-check: a second model reviews the spec and the finished diff — Claude checks ChatGPT and the other way round.
AI and agents: what can already be built with them. The history of technology and how many leaps fit into one lifetime — there is a piece of stuff about that.
Three cases: the problem, my role and the result.
A payments platform for merchants: accepting payments, merchant dashboards, roles, integrations with payment providers in different countries.
A subscription VPN service in Telegram: a bot, a Mini App, connecting in a couple of steps, servers and operations.
Pipelines in Mage.ai, Superset dashboards with role-based access, and my own mobile shell over Superset that opens the same dashboards on a phone.
Small services, experiments, stubs and collections — no common theme.
Things I make for myself and out of curiosity. Each one lives at its own link: open it, try it, share it.
My main stack is JavaScript and TypeScript.
APIs, business logic, event processing and integrations
Web interfaces, admin panels and Telegram Mini Apps
Storage, ETL pipelines and BI dashboards
Bots, webhooks, data collection and browser automation
Deployment, monitoring and logs
Unit, integration and end-to-end tests
Implementation, specifications and review with a second model
TypeScript, Python and database work
Telegram works best.