AI researcher, applied AI engineer or AI product engineer
Three profiles, three different searches. A comparison table and the one question that tells you which hire your problem calls for.
Read the guide →Guides
How to hire AI researchers and AI engineers
Where each profile is found, when it is reachable, how to evaluate it and what an offer has to contain. Eleven guides, written from our own index of the published AI research and the public records that say when a doctorate ends.
The short answer
Hiring in AI fails on supply, not on process. The people who do the work are employed, publish under no job title, and are reachable for a short window around the end of a doctorate or a contract. Four in five sit somewhere a job board cannot see them.
Which changes the method. Search what someone published or built rather than what their title says, rank researchers on first authorship at a reviewed venue and engineers on public code, then plan the approach around the window instead of against it. These eleven guides say how, profile by profile, measured on our index of 1 048 947 researchers.
Which guide to read
The three profiles behind the acronym need three different searches, so start from the one you are hiring.
- You do not know which profile you need. That is the common case, and getting it wrong costs about a year. Start with AI researcher, applied AI engineer or AI product engineer, which compares the three on one table and gives the question that decides.
- You need an answer to an open question. Somebody who can tell you whether the capability can be built at all. Read how to find and hire AI researchers.
- The method exists and your problem is scale. Your data, your latency budget, your inference bill. Read how to hire an applied AI engineer.
- The model works and the product does not. Agents, retrieval, evals, the cost and latency work nobody demos. Read how to find and hire AI engineers.
- You would rather hand the search over. Read the best AI recruitment agencies, compared, which puts four firms side by side, our own included, on specialism, sourcing, fee and geography.
Why the figures here change
Every number in these guides is counted on our own index rather than quoted from a report, and each one carries the population it was computed on. The index is rebuilt nightly from arXiv, OpenAlex, ORCID, DBLP and Hugging Face, and the pages are refreshed every week, so a figure you read here is the figure we are working from. The method page says what counts as a researcher and where the numbers stop being reliable, and the open data page publishes the aggregates in CSV and JSON under CC BY 4.0.
AI research job titles, decoded
Research scientist, applied scientist, member of technical staff: what each title means, what to check before believing one, and which to advertise.
Read the guide →How to find and hire AI researchers
Search the record by subject, reach people six to nine months before the defence, evaluate with a research talk, and write an offer a lab cannot.
Read the guide →How to find and hire AI engineers
Where AI engineers come from, where to look in order of yield, and how to test one in an afternoon.
Read the guide →How to hire an applied AI engineer
Two routes into applied AI work, the twelve to thirty month window after a move into industry, and how to test a candidate in half a day.
Read the guide →Hiring NeurIPS, ICML and ICLR authors
What an accepted paper at NeurIPS, ICML, ICLR or CVPR actually tells you about a candidate, how many people hold one, and when to reach them.
Read the guide →Where AI researchers are, and when they come free
Where AI researchers sit by country and how many finish each year, with the door to one report per country and per subject.
Read the guide →The best AI recruitment agencies, compared
Four recruiting agencies for AI researchers and AI engineers, what each is built for, how they charge, and the questions to ask before you sign.
Read the guide →How long an AI doctorate takes, and when to reach the people finishing one
The years between a first paper and a defence, measured by subject, and what that arithmetic tells you about when to write.
Read the guide →The best tools to find AI researchers, compared
Five tools to find AI researchers, AI engineers and software engineers, from LinkedIn to GitHub, judged on what they read, whether they merge duplicates and whether they know when someone becomes available.
Read the guide →How to hire AI research interns for summer 2027
Where the AI doctoral students are, when to start for a summer internship, and what an offer has to contain to win one.
Read the guide →The index itself
Reading about it is the slow way.
Search any subject in your own words, see who leads the work on it, and hear months ahead when they come free. Or hand the search to us and get a shortlist back.