I build the boring AI systems that actually save time.

I've been building for the web since 2007. Custom WordPress themes, a 125-site multisite network, enterprise Webflow, and the analytics stacks that tell you whether any of it worked.

A few years ago the interesting problem moved. How to build the page stopped being the hard part. The hard part is working out which of a team's workflows is quietly eating hundreds of hours a year, and whether you can hand any of it to a model without regretting it.

So that's where most of my time goes now. Find the process nobody enjoys. Scope it small, build it, then go back and check whether it saved the time it was supposed to. Then pick the next one.

I still build the front end it lives on. Half the reason these things fail is that nobody can figure out how to use them.