I work in enterprise IT — a mix of hands-on support, infrastructure administration, and building small tools that remove repetitive manual work from day-to-day operations. My background spans large public-sector environments and smaller organizations, which means I've had to be equally comfortable following established process and building things from scratch when no process existed yet.
Most of what's below started the same way: someone was doing a task by hand, on a schedule, and it was slow, error-prone, or both. I like finding the smallest reliable fix.
Replaced a manual, once-a-week process of logging into multiple portals to download financial statements with a scheduled script that connects via FTP, retrieves new files automatically, and sorts them into the right destination folders. Cut a recurring hour-plus task down to a few minutes of exception-checking.
Built a pipeline that pulls text out of scanned documents with OCR, parses the relevant fields, and cross-checks them against internal records to flag mismatches for review. Turned a manual line-by-line comparison into an automated first pass that only surfaces the exceptions worth a human's attention.
A lightweight matching tool that compares transaction-level data across two independent sources and reports discrepancies — amounts, dates, or missing entries — instead of relying on someone manually reconciling spreadsheets side by side.
In a large public-sector device-management environment, rollout readiness across sites was tracked by hand from several disconnected systems. Built Python-based reporting that pulled data from inventory and mobile device management platforms, and turned it into Power BI dashboards showing which sites were ready for a phased hardware or software rollout and which weren't — replacing manual cross-referencing with a single source of truth the team could check before every phase.
A school needed student and staff records reshaped into a third-party single sign-on platform's required format so accounts and class rosters would provision automatically. Wrote a script that cleaned and remapped the source data to spec, replacing a manual account-and-roster setup process with an automated one and eliminating the data entry that came with every term's changes.