I build the unglamorous software a business runs on.

I spent years building software for companies of every size before starting this. The pattern that kept repeating: an inefficiency costing a business thousands of dollars a month, which would take about a week to fix, and which nobody was fixing — because the firms who could afford to do it were chasing larger contracts.

A technician spending twenty minutes turning a voice note into an invoice. A supervisor answering the same lookup question forty times a day. A crew retyping data that already exists, into a system that already has it, because the two systems have never been introduced.

That’s the work. Some of it is small — a text-message tool built in a few days for a few hundred dollars. Some of it is a multi-month system that fifty people depend on to get paid, with offline capture, a review portal and a delivery pipeline that cannot lose a job. The engineering standard doesn’t change between the two.

I build with AI-assisted tooling every day, and I’m direct about that: it is a real part of why one person can ship and operate systems at this scale. It is also not a substitute for knowing what a durable queue is, or why row-level security belongs in the database rather than the UI. The judgement is still the job.