Sydney / AEST — Open to work

Built to be
understood.

iOS and full-stack, from Sydney. Master of IT at UTS, graduating July 2027. I came to software from film, and it shows in how I build.

See the work ↓

// Selected work — 04 entries

Featured Projects

Four projects, each shipped under a deadline with other people in the room.

CookPilot home screen Ingredients identified from a fridge photo Recipes ranked by what you already have Hands-free step-by-step cooking view iOS · 1st place · ICON x Lyra 2026 CookPilot

Look at your fridge through Ray-Ban Meta glasses and it tells you what you can actually cook tonight, read into your ear step by step.

The walkable 3D gallery with AI-generated images visionOS · Apple Foundation Program 2026 Visual Eyes

Asks who you want to become, then builds a five-room museum of that future and makes you walk through it.

8 of 8 attacks blocked Swift SwiftUI Python MCP MCP · OpenAI Codex Hackathon LearnGuard

A gate that makes a coding assistant earn its permissions. The red-team suite blocks 8 of 8 attacks while letting legitimate actions through.

VCF in, ranked candidates out Python FastAPI SQLite Playwright FastAPI · HSIL, Harvard T.H. Chan VaxAgent

Tumour mutation data in, ranked neoantigen candidates out, with every step traceable back to the file it came from.

// Capabilities

What I can build

Six things I have actually shipped, not a list of tutorials.

iOS apps

Swift, SwiftUI, on-device speech recognition

Spatial

RealityKit, visionOS, USDZ scene generation

Backends

Python, FastAPI, SQLite, streaming progress

Wearables

Meta Wearables DAT on Ray-Ban Meta glasses

AI plumbing

OpenAI, MCP servers, tool-permission gating

Shipping fast

Hackathons, tight deadlines, other people in the room

// How I work

From idea to shipped

Learned the hard way, mostly on CookPilot.

01 Find what is actually defensible. Partway through CookPilot we found that every layer of the idea had already shipped somewhere — eight apps do fridge-photo-to-recipe, and Meta's own assistant reads steps aloud. Deciding what was still ours mattered more than anything we built.
02 Prototype the risky part first, before anything that is merely fiddly.
03 Test it on someone with no context. Ninety seconds, no explaining from me. This is the part film school taught me, and it is the part most projects skip.
04 Ship it, then cut whatever nobody used.

// Away from the keyboard

I climb, swim, run, and get out of the city most weekends. Last winter I cycled from Seoul to Busan — seven hours a day, several days running — which is the closest thing I have found to a hackathon that lasts a week.

I am learning to write and think in English rather than translating into it. This site is part of that.