I build the parts of a system that have to keep working when the traffic is real: data planes, control planes, the tooling around them, and the backends behind them.
Most recently I built CC Usage Stats, a macOS menubar app, without ever having built a Mac app before. I didn't learn Swift for it. I ran the project the way I used to run engineering teams — specs, tests, independent review and a definition of done — and agents wrote the code. Seven months on it has shipped 82 releases.
Writing software since the late nineties; the last twelve-plus years in distributed systems and backend infrastructure — much of it in Go, most of it in networking, observability and databases. Since 2025 I work AI-first: specs and tests written by me, implementation by agents, gates no model can talk its way past. By day I am a senior backend engineer at Spamhaus, where I co-wrote the AI usage policy and built the internal Claude Code plugin marketplace.
Work
-
CC Usage StatsmacOS menubar app
How much Claude do I have left? A menubar gauge for Claude and Codex rate limits, with alerts before you hit the wall and a read-only MCP server so agents can check their own headroom before a long task. Open source, MIT.
Writing
-
The best skill for building a Mac app was managing people
I had never built a Mac app. Seven months and 82 releases later, the skill that mattered wasn't Swift: it was running a team, except this team was agents.
Background
- Spamhaus Technology — senior backend engineer; internal advocate for AI-first development
- Form3, Monzo, Divido — backend and infrastructure engineering in payments and banking
- BrightLocal — head of infrastructure
- Ciklum, Stuzo, Sterling Hoffman — earlier engineering roles
- MSc Computer Science, University of Liverpool
Contact
I occasionally take on infrastructure problems alongside my day job, and I am always glad to talk about data planes, agentic development in codebases that cannot afford to be wrong, or anything above. Email is best: hello@dmytro-vovk.com.