Product software, on-device ML, and systems engineering.

A private iPhone and Apple Watch product built to turn live therapy context into forward-looking, reviewable decisions around meals, corrections, and exercise.

• Built an on-device prediction experience that makes live therapy data useful before action, not just after the fact.
• Designed meal and activity what-if flows, insight surfaces, and bounded personalization so the product becomes more useful as history accumulates.
• Shipped a shared state layer reused across the iPhone app, widgets, Live Activity, and Apple Watch instead of rebuilding logic per surface.
• Combined product design, ML integration, persistence, and cross-device runtime architecture into one coherent system.

A long-running systems project built from the metal up: boot flow, kernel, memory management, executable loading, native desktop surfaces, and local inference running inside an OS I wrote from scratch.

• Spans the full stack from early boot and kernel bring-up through scheduler, VFS, section mapping, and user-mode runtime.
• Implements the kind of systems work usually split across multiple teams: paging, file-backed memory, executable loading, process plumbing, and native debugging.
• Brings local LLM inference directly into the platform with GGUF-based runtime components and agent-style workflows.
• Pairs low-level infrastructure with visible product surfaces including a desktop shell, Finder-style tools, and Task Manager-style diagnostics.

Supporting platform work that pushed the operating system beyond a kernel demo and into something that behaves like a usable software platform.

• Built driver-backed IPv4 networking with ICMP, UDP, and TCP support, plus TLS client tooling for secure outbound requests.
• Implemented windowing, desktop interaction patterns, and first-party native tools that make the platform feel usable rather than academic.
• Used these layers as a proving ground for debugging, integration, and reliability work across kernel, runtime, drivers, and UI.

A retired macOS menu bar utility formerly distributed through the App Store for viewing live Dexcom glucose data without opening a separate app.

• Pulls glucose data from Dexcom Share and shows current value and trend at a glance.
• Built from a personal need to reduce context switching while working.
• No longer maintained and retired from the App Store in 2024; included here as a concise project reference.