U.S. AI jobs could go unfilled by 2027
BainThe CHRO’s
AI Talent Blueprint
AI engineering job postings year over year
LinkedIn, 20257% of technical job postings ask for AI skills. Under 1% of U.S. workers have them.
LinkedIn, 2025Employers plan to recruit AI tool builders
WEFBanks and manufacturers now recruit the same AI engineers as the software firms do. They draw from one pool, and it is growing more slowly than the job postings.
*Figures are third-party research findings published by the sources cited and are presented for informational purposes only. They describe general labor-market conditions as of the dates of those reports, are not Carver Edison data or results, and may not reflect your organization’s experience.
Set the business mandate
Name the business outcome before you name a single role. Once you know the outcome, you can see which capabilities you have to own and which you can rent.
Map the capabilities
AI work splits across four layers, and most teams need people in all of them. Pick the roles your outcome requires and leave the rest.
Redesign the work
A new title on an old workflow changes nothing. Sort the work into what you automate and what your people keep, then write the role from what is left.
Redesign the workflow first. The role is the judgment you still need a person to own.
Choose how to access talent
Some of these roles you should own outright. Others you can grow from people already on your team, or rent from a partner while you build.
Differentiating capability you need to own.
Adjacent talent can grow into this role.
Capacity you can rent while you build.
Build the candidate opportunity
The engineers you want are choosing between offers. They weigh the problem, the people they will learn from, and their share of what they build.
Your one-page blueprint
Take this into the room with your CTO and your business leads. It is the one page you can all mark up together.
This resource is provided for informational and educational purposes only. It is not legal, tax, investment, or compensation advice. Outputs reflect the inputs you enter and should be reviewed by your own personnel and advisors before use.
LinkedIn defines AI engineering as the technical skills that build and deploy AI systems: machine learning, predictive modeling, and neural networks.