Now

Work

At UST, working on a large-scale mobile and web product - React Native on mobile, React on web, with a lot of cross-platform surface area to keep consistent. Focused on architecture, performance, and making sure the team can move fast without the codebase turning into a problem.

Building

JIB - a personal JARVIS-style iOS app with an audience of exactly one: me. Two local AI brains (Apple Foundation Models on-device, Gemma 4 12B on my MacBook via LM Studio), and a generative UI where the model composes native SwiftUI screens instead of chat bubbles. Currently exploring Siri integration through App Intents.

Learning

• On-device AI as a first-class platform - Apple Foundation Models, App Intents, and what changes when inference is free and private.
• Running local models seriously - Gemma 4 12B in LM Studio on my Mac, tunneled to my phone. Inference lives on the Mac; the homelab self-hosts everything around it.
• Agentic coding workflows - orchestrating Claude Code with advisor and verifier subagents instead of treating AI as autocomplete.

Reading

• X feeds and tech newsletters - staying close to what's shipping in the AI and frontend space.

Outside work

Family time, tinkering with the homelab, and playing games on the PS5. The things that remind me there's a world outside the terminal.