Sample report: an AI code audit of my own repo
A finished AI code audit of my own directory-engine repo. 50,000 lines, one week, mostly AI. Ten findings down to the file and line, with hour estimates and a fix order. Published as it came out.
NOTES & ESSAYS
Writing
Posts on AI, agents, and the practical edges of building with them. Older work originally published on Medium is linked out below; newer writing lives here.
-
-
It works on your laptop and dies in production. What's actually wrong with the app you vibe coded
The code the AI wrote is mostly fine. What breaks is the seams: environment, network, auth, data, deploy, load. Seven ways it shows up, what's underneath, and what fixing each one tends to take.
-
Most agent demos won't survive a real codebase — here's what does
The demo is the agent answering a question. Production is the gateway, the pipeline, and the auth service nobody saw coming, and that's the part that decides whether an AI feature survives.
-
· Medium
AI use cases: personalized podcasts from your own sources
An app idea that turns the articles, papers, and videos you can’t keep up with into AI-generated audio for listening on the move.
-
· Medium
Running n8n Locally with Ollama: A Cost-Effective Automation Solution
If you have something to build or fix, send me a few sentences and I'll reply within two business days.
Get in touchRSS feed · all writing across venues