Look across the careers pages of the leading AI labs and a pattern emerges. The race for talent has narrowed to a handful of roles that almost every frontier company is chasing at once.
The shortlist
Top of the list are research and machine learning engineers who build and train models. Close behind are forward deployed engineers, who embed inside customer operations to make the technology actually work in the wild, a role that barely existed a few years ago.
The list continues with data and evaluation specialists who measure model quality, infrastructure engineers who keep vast clusters running, and a fast growing cohort of safety, governance and policy staff.
Why it matters for everyone else
The convergence sets the going rate for AI talent across the whole economy. With 1.6 million AI roles open globally against roughly 518,000 qualified candidates, the labs' shopping list is effectively pricing AI skills for every other employer too.
