A self-learner.
A leader. A gamer.
A software developer.

I started tinkering with computers when I was fourteen — installing Windows, figuring out drivers, seeing what a machine would let me do to it. Most of it was in service of games. I wasn't trying to become anything yet. I was just curious about what was underneath.

I took the long way round. I started out in medicine, then switched to ICS so I could study computers properly. Programming didn't really click until university — and once it did, I kept my GPA above 3 on purpose, because I knew that number was what would get me through the door.

I've been a curious person from the start, always looking for answers and for a way to actually build the thing I was curious about. That's still what pushes me to keep learning.

The problems I enjoy most are the messy, real-world ones — a process that currently lives in someone's head, or across a few spreadsheets, that needs to become software. I like integration: making one system actually talk to another, correctly. And I like owning something end to end rather than one slice of it.

Mostly, though, I like working on things that are new to me. Even when I already know a tool, I want to know where it's headed next — which means I'm rarely "finished" with any of this. I've deployed Node services on EC2, made calls and sent SMS through Twilio, and integrated Google Workspace, Apollo, SendGrid and Resend. When a project needs something new, I learn it to the depth the job requires.

A good example of how that plays out is a project called Money Diary . People kept telling me to just keep a written diary to track my spending, but I knew that would turn into a pile of notes I'd never want to search through or summarise — so I built it instead. Along the way I learned how to integrate cheap AI solutions like OpenRouter, and how to think about context: what a model actually needs to be told to be useful across a conversation.

More recently that has meant generative media. AIWalls is an independent AI image lab: eighteen local experiments, each with the prompt, model, sampler and seed attached, plus a model bench and a same-prompt comparison. Voice Clone Studio keeps voice cloning and speech synthesis on my own GPU instead of someone else's API. RGB Studio is the other side of the same instinct: a Rust service and a desktop app that control a piece of hardware without a cloud account.

What I want the people I work with to come away thinking is straightforward: that I'm a good, self-improving person to have on a project, that I respond quickly, that I'm real and trustworthy, and that I treat whatever I'm building for someone else as if it were my own.