Getting Started

Hi, I’m Brad, and I’ve started Robot Brains as my AI Lab.

I’ve been writing software as a professional since 1998, working for over a decade at tech darlings like Xero and Atlassian. In 2015, I designed and productionised my first AI product. It was well before the current hype cycle—simple by today’s standards (logistic regression and scikit-learn)—but it works amazingly and is still crushing it today.

The current landscape of LLMs is quite different, and I’m keen to explore how I can use them in real use cases.

Projects in the Lab

I have two projects running at the moment:

AI Onboarding Agent

I’m exploring how to utilise an LLM to gather structured information from a customer—client onboarding, for example. The challenge is seeing if a single developer can create a killer product by using AI tools for everything: product design, timelines, UX, and UI.

Legacy Desktop Bridge

I’m investigating whether I can help a friend automate flows in their Windows desktop application using AI. The challenge is that this is a legacy system, so connecting to AI needs to be “in-process” rather than the standard HTTP request/response model. And because it’s a secure environment, I’m figuring out how data stays in the “walled garden” using anonymisation.

Why this Journal exists

The goal of the lab is to learn through doing, and maybe create something great along the way.

I’ll be using this site to share what works and—more importantly—what doesn’t, as I move these experiments out of the sandbox and into real environments. Progress won’t be fast (evenings and weekends), but the challenges are real.

— Brad

Disclaimer: This project is conducted in my personal time. All thoughts and opinions expressed here are solely my own and do not represent any current or former employers.