I put AI features into apps people actually use, and get them through App Review.

Open to senior mobile roles and contract work — remote

What I care about is the part after the code runs. Anyone can get a feature working on a simulator. Getting it through App Review is a different job. So is keeping it stable on old devices and bad networks. So is being able to roll it back when something goes wrong. On mobile you cannot just push a fix.

Current engineering interests

What App Review will and will not accept from an AI feature
Where inference belongs — on the device, or worth the round trip
Agent tooling and browser automation
Designing what an assistant should refuse, before what it should answer
Clean Architecture and BLoC in codebases that already have users
Real-time systems: WebSockets, chat and presence

Now

NOW / 01Open to senior mobile roles and contract work, remote.
NOW / 02Maintaining browser-execution-runtime — a provider-agnostic browser runtime for AI agents.
NOW / 03Five applications shipped to the App Store this year, four of them AI.