About OWNML
Most ML projects don't fail on technology. They fail because nobody agreed what to predict, who acts on it, or what getting it wrong costs. By the time that's obvious, the model's already built.
I'm Louis Dorard, and I started OWNML to fix that.
Where I come from
I've been designing and delivering ML systems since 2014, first as an independent consultant. The pattern showed up early and never really changed: ML projects rarely fail on modeling. They fail on design and alignment, before anyone touches data.
That's what the Machine Learning Canvas is for. I built the first version in 2015 — a one-page way for business, data and engineering people to design an ML system together, end to end.
From canvas to method
Over ten years, the canvas has been downloaded by 15,000+ practitioners, through consulting, training and conferences. That field experience is what the rest of OWNML is built from: a structured way to design, validate and ship ML systems that create real business impact, not just a working model.
What I focus on
Models trained on your own organization’s data
Tabular, operational use cases: classification and regression
The whole path from prediction to decision to impact, not just the model
No hype, no tool-first thinking. A clear way to make an ML project work.
