Healthcare IT systems specialist in Cambridge, Ontario. I turn messy operational work into repeatable PowerShell and Python automation, and I build local-first utilities that keep private data on the machine they run on.
Website · LinkedIn · Full portfolio index
Download for Windows · View project · Report a problem
Plan when your Codex day starts.
Your 5-hour window only begins when you actually use Codex, so a window that ends overnight leaves the next reset sitting idle until you come back. UsageLoop starts it on a schedule you set: one time, separate weekday and weekend times, or a different time for every day. It keeps windows rolling through the day and pauses overnight. Local-first: no API key, telemetry, or cloud account. It does not add quota or bypass limits. Windows, with a Linux beta in testing.
Try it in your browser · Windows and Linux downloads
Cleans secrets, hostnames, and personal details out of scripts, logs, and prompts before you share them. No backend and no account. The part I use most is the PowerShell-aware Portfolio-code mode: it swaps organization-specific terms for readable generic ones, so sanitized code is still worth reading.
Download for Windows · See it first
Private personal finance. Import your own bank statements and get a straight answer about where the money went. No account, no subscription, and the database is a file on your computer. Tauri and Rust shell, React and TypeScript interface, packaged Python finance engine, SQLite, close to 1,400 tests.
Sanitized versions of real healthcare IT automation: onboarding, password remediation, Windows 11 device replacement, O365 migration support, workforce identity migration, and code review utilities. The private data is stripped and the workflow patterns are intact.
Wedding 50/50 Draw · live site. React and TypeScript on a Google Apps Script and Sheets backend, built for a real event. Live totals, confirmation emails, printable draw slips, tests, and Pages deployment.
benthompsondev.github.io · live site. My personal site.
The shape is the same in almost everything here:
input -> validate -> plan -> review -> act or simulate -> log -> verify
I care about validation, logs, rollback paths, and output another person can actually review. I am heading toward automation-heavy DevOps, cloud automation, and practical software engineering.
The full portfolio index has everything else, including older teaching and course work.
Questions about any of it, or want to talk shop? Find me on LinkedIn.