I like building software that sits close to real people: developer tools, small product surfaces, weird side projects, accessibility work, and the glue that helps teams ship with confidence.
Most of my favorite work lives in the messy middle: turning a confusing workflow into something clearer, making an API or CLI feel obvious, writing the docs I wish existed, or polishing the little details that make software feel trustworthy.
Lately, I've been especially interested in practical agent workflows: how AI tools produce changes, how developers review and apply them, and what it takes to make that loop trustworthy enough for real work.
- Developer tools that feel obvious once you use them
- APIs, SDKs, CLIs, and docs that reduce support load instead of creating it
- Accessibility as product quality, not a checkbox
- Small teams with high trust and strong user feedback loops
- AI-assisted engineering workflows that ship real code
- OrbitDock - Mission control for AI coding agents across Claude Code and Codex CLI.
- Vizzly - Visual testing and review for product teams.
- Pitstop - Vehicle telemetry, maintenance, and track-side tooling.
- Tornadic - Crowd-sourced tornado data and storm history.
- MoonBun - Rabbit care tracking for households, health records, reminders, and boarding handoffs.
If any of this overlaps with what you're building, say hi.