I got into analytics because I kept watching good data go unused.

My first three years of work were in support and billing operations — 200+ retail accounts at a POS company, then billing and compliance at a Y Combinator–backed studio. Unglamorous work, and the best analytics education I could have had, because it put me next to the data while it was still messy and still mattered.

The pattern I kept hitting was always the same. The numbers existed. Somebody had even built a report. But nothing in the system ever acted on them — the reconciliation was still manual, the billing still depended on someone remembering, the insight died in a meeting. The gap was never analysis. It was the distance between an insight and a working process.

So I started closing that distance myself. A fractional billing system that put six figures a month on rails. An HR agent that generated its own documents and fed worklogs straight into billing. Slack integrations that handled compliance chasing. None of it was in my job description, which is rather the point.

Getting the formal half

What I didn't have was rigour. I could build the system; I couldn't always prove the answer was right. That's why I'm at Smith School of Business at Queen's University doing a Master of Management Analytics with a Data Science major — optimization, statistical inference, machine learning, simulation. The half that lets you defend a recommendation rather than just ship it.

The coursework has been the most enjoyable part of it: a mixed-integer program that prices what menu variety costs a hospital, a market-value model across 4,056 footballer-seasons, a top-20% Kaggle finish. Not because they're clever, but because each one ends in a decision somebody could actually take.

Running something of my own

In the middle of all this I started Hopefield, a home-textiles brand selling on Amazon FBA and Shopify. Partly because I wanted the challenge, and partly for a reason that turned out to matter more than I expected: I wanted a business where I owned every number and couldn't hide behind anybody else's dashboard.

It has been the sharpest teacher of the three. When you write a linear program to decide how to kit your own inventory, and the leftover stock is your own money sitting in a warehouse, you stop treating model assumptions as an academic formality.

How I work

• Start from the decision, not the dataset. If I can't say what someone will do differently because of the analysis, I haven't found the problem yet.
• State the limitations out loud. A model with known, published limits is worth more than one presented as truth.
• An absurd result is a gift. The dangerous bug is the one that produces a plausible number. I check outputs against the real world, not just against a validation score.
• Automate where two teams maintain the same fact. That seam is almost always where the highest-value automation is hiding.

Outside the work

I grow food hydroponically in two tower gardens and log it obsessively, which is either a hobby or a longitudinal experiment depending on who's asking. I'm based in Kitchener–Waterloo and I'm open to analytics roles across the GTA and Southwestern Ontario.