Hello, I'm FROWNINGdev.

I am 🎥 Computer Vision Engineer and 🧰 Developer-tools maintainer ( 📰 featured in Django News #347 and PyCoder's Weekly #746 )

Welcome to my page; on my Github, you can find:

• Open-source developer tools written and maintained by me
• Production smart-city systems I build at work — ANPR, video analytics, SCADA
• Notes on Django internals, static analysis and camera infrastructure

I am open to new opportunities.

_{Generated by jstrieb/github-stats rather than the usual card service, because most of my work lives in private production repositories — a public-only counter would show a fraction of it.}

I ship production smart-city systems end to end: from the GPU detection pipeline to the dashboard an operator actually uses. ANPR / license-plate recognition, real-time video analytics, PTZ camera infrastructure, and industrial SCADA.

The constraint that shapes most of my choices is that these run on real cameras, in real weather, in front of people whose job depends on them not lying.

django-orm-lens — reads a Django project's model graph without booting Django: no import, no database, no venv of the target project. regex + ast answers "what points at User?" in ~15ms instead of a 2–5s Django boot.

One core, three distributions — a Python CLI, a VS Code extension, and an MCP server for AI coding agents. The TypeScript and Python parsers are pinned to identical output by a shared golden fixture, so the editor sidebar and the agent can never disagree about a schema.

📰 Featured in Django News #347 and PyCoder's Weekly #746 · listed on the official Model Context Protocol Registry.

_{Most of these are private production repositories — which is exactly why the stats above are generated rather than counted from public activity.}