Tools and libraries I built in 2026, client work from my agency years, and an archive of earlier projects.
From 2021 to 2026 I worked on client product teams at Gorilla Logic (Panasonic Well, Cyngn, HIMSS, The Pro's Closet) and on automation projects for my own clients. That code is private, so the recent section below is personal work from 2026, and the blog has case studies from the job.
2026 tools and libraries
Sits between an app and an LLM API: masks names, SSNs, card numbers and other PII with typed placeholders, then restores the real values in the reply and tool calls. Per-category mask, block or allow policy, an audit log that never stores values, and zero runtime dependencies.
Benchmarks TypeSafe's Jev against Claude Haiku 4.5 and Opus 5 on 72 labeled decisions, with the same questions and scoring code for every model. Jev scored 94.4% at ~185 ms and ~$0.02 per 1k calls; the report also covers calibration (ECE, Brier).
Shows disk usage as a sorted tree, scanning in parallel. Scans 153,000 files in 0.27 s, against 0.64 s for du.
Finds duplicate files by size, then BLAKE3 hash, and reports the space they waste without deleting anything. Found 44,246 duplicates across 151,197 files in 2.5 s.
Books tennis courts the moment the reservation window opens at midnight, in about 500 ms, 12x faster than the browser-automation version it replaced. Runs on cron and emails a confirmation.
Agency projects, 2018 to 2020
Landing page for American Express card holders that personalizes its content from customer data.
2021 learning projects, demos offline
