Leading AI Initiatives
Who to Hire First (Not a Researcher)
6 min+55 XPStep 1/10
The hiring order that works
The default sequence for a company adopting AI: first, a product-minded engineer — strong software skills plus enough ML fluency to wield APIs, build evals, and ship. Then, product ownership — someone accountable for the metric the feature is supposed to move.
Researchers and PhDs come later, if your roadmap graduates to capabilities nobody rents — most never do. The inverse order (hire a research star, find a problem later) produces papers, not revenue.
insight
The scarce skill in applied AI isn't training models — it's evaluation and integration. Hire the person who builds the eval harness and ships the feature, not the one with the most citations.