What I build outside work.
Two AI-native products I designed and shipped end to end, directing AI coding agents: a mental-health platform for Australians, and the device-review tool I use to ship it.
FindPsych
AI-native matching of patients to psychologists, with sensitive health data handled by design.
The findpsych.com.au prototype matches patients with psychologists in Victoria through an AI-native chat interface and research-backed patient–psychologist preference matching, with rigorous security and compliance protocols governing sensitive data.
- Patient intake is live, against synthetic practitioner profiles: describe what you need in your own words, or filter by support, format, location, language and time
- Matching built on research-backed patient–psychologist preferences, not just a postcode
- Compliance as architecture: sensitive data handling and AHPRA requirements designed in from the start
- Map and list views with transparent fees, suburb-approximate pins and no ratings
Kview
Device review for people who build with AI coding agents.
One URL rendered across phones, tablets, laptops and ultrawide at once, each a genuinely emulated device rather than a narrow desktop window. Capture what’s wrong, mark it up, and hand your agent a review it can act on.
- Real device emulation through the Chrome DevTools Protocol: viewport, pixel ratio, touch, user agent and client hints
- One-click captures at true device resolution, with numbered marks your notes can point to
- Exports the whole review as markdown, each screenshot tagged with its exact viewport, pixel ratio and URL, ready for Claude Code, Codex or Cursor
- Works on localhost and deployed sites alike. This site and FindPsych are reviewed with it
