DistanceDefender
A macOS utility that watches your screen distance and posture, then nudges you to look away before eye strain sets in. All processing happens on device — no camera frame ever leaves the Mac.
macOS · iOS · on-device AI
I'm Nathan. I build native macOS and iOS software with privacy-first architecture — and the on-device AI that runs inside it. Welcome in; here's what I've been working on.
everything above stays inside this machine
About
I build for the Apple ecosystem, and I lead the teams that ship it.
My work runs the full length of a product — system architecture and technical documentation through to App Store Connect deployment and compliance — with a consistent bias toward local-first designs that do their work on the user's machine rather than someone else's.
Fluid, modern experiences for macOS and iOS built with SwiftUI, WidgetKit, and liquid-glass interface design.
Steering technical teams and complex projects from conception through deployment, holding velocity as Scrum Master.
Secure, local-first applications built on offline-only entitlements and on-device processing.
Architecture and technical documentation through App Store Connect submission, review, and compliance.
Pushing what Apple Silicon can do directly — ComfyUI with MPS acceleration, local LLMs, no round trip.
Context-aware systems built on GraphRAG, LangChain, and Model Context Protocol servers.
Tools aimed at real problems: health-focused monitoring utilities and educational financial literacy platforms.
The intersection of software, energy infrastructure demand, and financial data analysis.
Selected work
Things I've built, and what they're for.
A macOS utility that watches your screen distance and posture, then nudges you to look away before eye strain sets in. All processing happens on device — no camera frame ever leaves the Mac.
An educational platform that teaches financial fundamentals through interactive scenarios instead of lectures. Learners practise consequential decisions in a sandbox before they face them for real.
A local-first workspace for running foundation models on Apple Silicon with MPS acceleration and no cloud dependency. Built to make on-device inference practical enough for everyday tools, not just demos.