Geospatial at the core
Field borders, management zones, machine-ready prescription files.
Complex software · Applied AI/ML · Technical leadership
We build the software other shops call out of scope — data pipelines, numerical models, and the products wrapped around them. Twenty years in, we still do our best work where software meets math, science, and messy real-world constraints.
What we do
Most engagements begin in one area and draw on another. That's fine — the same people work across all three.
Data pipelines, optimization engines, and scientific software built around the real constraints. If the work needs math, domain reasoning, and careful engineering in the same room, send it here.
Learn more →02Machine learning and automation scoped to the decision they improve, with evaluation and deployment built in. Useful systems, not demonstrations.
Learn more →03Fractional CTO work: architecture calls, hiring, vendor decisions, and a roadmap that survives contact with the budget. Senior judgment, sized to the work.
Learn more →Our own product
YouFarm is a precision-agriculture platform we designed, built, and operate — from soil tests to variable-rate prescriptions used each season on working farms. It brings our geospatial, optimization, and product work together in one production system.
Field borders, management zones, machine-ready prescription files.
Inputs allocated across a whole farm under real cost and agronomy constraints.
Data model, math, maps, and interface — one small team, years of iterations.
About us
We're a small team of software engineers and scientists based in Western Canada, with decades of experience across AI, applied research, and technology leadership. We like problems where the answer isn't in a tutorial.
Our aim is unglamorous and specific: deliver automation, modelling, and analysis that still works in two years. We'll tell you when a model isn't the answer.
New ideas are welcome; unverified ones aren't. We show the working.
You get what you need and nothing more — including when that's less than you asked for.
Code that reads well a year later. Documents someone can actually follow.