I did the operational work before I started automating it

Before I wrote software, I worked in sales, customer service, financial transactions, inventory, and small-business operations. I have been the person copying data between systems, chasing down a duplicate record, and finding out at the worst moment that a "simple" process had six exceptions nobody wrote down.

That experience is why my software looks the way it does. I notice where a workflow actually fails — not where the diagram says it should — and I build for the operator who has to live with the result. Uncertainty becomes a review queue instead of a silent guess. Consequential actions wait for a person.

Today I build focused tools around that principle: Python pipelines for messy data, browser tools for repetitive research and form work, and internal applications that make cleanup and review explainable.

AI has a place in that work, but it is a layer, not the foundation. When a plain rule can do the job, I use the rule. When AI helps draft or summarize, the output stays editable, it is clear what data leaves the machine, and a person approves the result. I direct the work, test the behavior, and own what ships.

I'm open to freelance projects, contract engagements, and full-time roles where operational judgment and practical software belong together.