About

I started my career in Grenada, working as a meteorological observer and later as a forecaster. Over the years, I issued public forecasts, aviation products, marine advisories, and warnings under real operational pressure. Weather wasn’t theoretical — it affected flights, farmers, emergency managers, and everyday life.

But something kept bothering me.

So much of the work relied on fragile workflows: manual downloads, outdated scripts, disconnected systems. Forecasters were spending time fighting tools instead of thinking about the atmosphere.

I didn’t want to complain about it. I wanted to fix it.

I taught myself Python. At first, it was small things — automating downloads, cleaning datasets, plotting charts. Then databases. Then internal tools. Then web applications. Eventually, I realized I was no longer “just” a meteorologist using code — I was becoming a software engineer who understood weather deeply.

That realization led me to formally retrain in software development in New Zealand, where I worked as a Meteorological Software Developer and completed a Master of Software Development. There, I migrated legacy scientific code, automated model data ingestion, built ensemble visualizations, worked in HPC environments, and applied machine learning techniques to real weather datasets.

What ties all of this together is a simple idea:

Good weather decisions require good systems.

I care about building tools that are:

scientifically sound

operationally reliable

well-documented

and humane to use

I’m especially motivated by improving weather services in small and developing states, where the impact of better tools is immediate and meaningful.

Right now, I’m focused on the next step: machine learning for meteorology — learning how modern AI models can complement physical understanding, not replace it.

This site exists to document that journey in public.