About

I’ve been writing programs since 1985 and getting paid for it since 1997. These days I live in the Chicago area. What I work on now is on the home page, and the dates and details are in my resume. This page is the longer story.

How I got here

My first computer was an Elektronika B3-34, a Soviet programmable calculator. I liked writing programs for it, and I kept liking computers long before anyone around me thought they were cool.

My first paid job, in 1997, was writing software for a city council in eastern Ukraine, while I studied at the Ukrainian Engineering Pedagogics Academy. That was the era of FoxPro 2.6, Clipper, Clarion and Delphi. I still remember how to write all of them, even though nobody has asked in a while.

From 2001 to 2003 I wrote Linux kernel modules for billing systems. Then I moved to a ferroalloy plant, first as a developer and server administrator, and from 2005 as head of the IT department, with ten people looking after the plant’s systems.

In 2009 I took over development and IT at a cable broadcasting company. That job was as much physical network as software: replacing 50 km of copper with fiber, launching digital TV and DOCSIS internet, and merging three head stations into one.

From 2012 to 2017 I worked as a software and DevOps engineer at a payment processing company, on high-load systems and PCI DSS compliance.

In 2017 I joined OpticsPlanet’s office in Ukraine as a DevOps engineer. After two years there, the company moved me to its US headquarters, and I have been in the Chicago area since 2019.

Outside work

I have been swimming since I was six and still swim regularly. I speak Ukrainian, Russian and English.

About this site

The site is built with Hugo and hand-written templates, with no JavaScript framework and no external fonts or CDNs. The design is a re-creation of The Monospace Web by Oskar Wickstrom, set in JetBrains Mono. It is not a blog: the articles appear when I have something worth writing up.

Get in touch

I’m happy to hear from people working on similar problems, especially:

• Elasticsearch clusters that are slow, too big or too expensive
• HAProxy configuration and log pipelines
• Complex Varnish configurations
• Metrics platforms: Prometheus, VictoriaMetrics, OpenTelemetry
• Setting up LLM agents (MCP servers, Claude Code, Codex, OpenCode) for an engineering team

I’m open to consulting work and to conversations about interesting problems. Contact details are at the bottom of the home page.

Opinions on this site are my own.