I turn messy operations data into clear systems people can trust.

I like the parts of data work most people skip past — cleaning up a messy export, figuring out why two reports disagree, writing a prompt that actually extracts what you need instead of something almost-right.

Most of my time goes into taking unstructured information — service records, issue logs, freeform technician notes — and turning it into something structured enough to query, chart, and act on. That means SQL and Python day to day, and increasingly, LLMs doing the first pass so people don't have to read everything by hand. A mechanical engineering background means manufacturing and hardware failure data feels familiar rather than foreign.