Field note / posts
Why I went back to university
I had built useful things with AI assistance, but useful output was not the same as understanding the system underneath it.
A few years ago I tried to originate an advisory engagement for a startup using alternative data for credit scoring and revenue-based finance.
I could understand the financing proposition. I wrote the deck, arranged resources, and went into meetings with the founders. What I could not do was evaluate the technical claim at the centre of it. I did not know what evidence to request, what a working system should look like, or how far a demonstration was from something that could operate in production.
No money was lost. Time was, including time that belonged to other people. I left the engagement unhappy because I had been useful around the transaction but unable to test one of its main assumptions.
Building without understanding
I started with Python and Streamlit. One early workflow used on-device OCR to extract text from scanned invoices, passed that text to a language model to produce a reasonable JSON object, and sent the result into a self-hosted accounting system through its API.
That first version involved code I wrote and adapted from examples. Later, AI-assisted development made it possible to build much more: two App Store applications, an investment-research pipeline, and a content system that reached well beyond a prototype.
The tools worked. That did not mean I understood everything that happened between an instruction and the output. Vibe-coding shortened the distance between an idea and a working interface; it also made it easier to cross gaps in my knowledge without noticing them.
The reason for formal study
I do not need a computer-science degree to prove that a scanned invoice can become structured data. I do need stronger foundations if I want to judge technical claims, design systems that handle financial information, or build an AI-native advisory practice without treating the technology as a black box.
In 2026 I started the Master of Information Technology at the University of Western Australia, majoring in Artificial Intelligence.
The aim is not to rewrite my past as a software engineer. It is to become harder to mislead, more precise about what I build, and more useful on problems where finance and technology depend on each other.