What AI and machine learning hiring actually covers
It is rarely one profile. The brief gets sharper when you name the sub-area you are really hiring for:
- Generative AI and LLMs
- NLP
- Computer vision
- MLOps
- Applied and research ML
- Recommendation and ranking
- Speech
What good looks like
The strongest signal is production, not notebooks. Look for models that actually shipped, were monitored and were improved, and for people who reach for the simplest approach that solves the problem rather than the most fashionable one.
For applied roles, MLOps maturity matters as much as modelling: versioning, evaluation, deployment and monitoring are what keep AI working after launch.
How we hire for it
We start by separating the two very different needs behind "we need AI people": product ML that ships to users, and research that pushes the state of the art. Then we map specialists who have done that specific work and assess them on real problems, not keyword matches.
Because strong AI engineers are in high demand and rarely looking, most of this hiring is direct, discreet outreach, not job adverts.
In short
Name the sub-area, insist on shipped work, and weigh MLOps alongside modelling. That is how you hire AI people who deliver, not just interview well.