Hire a Research Engineer
The people who read the paper on Friday and have it running on Monday.
A small pool, and everyone is writing to them
- Retained recruiters.
- LinkedIn Recruiter seats.
- Keyword-matched resumes.
- Templated InMails.
A research engineer sits between the paper and the product: reproducing results, running experiments at a scale the paper never tried, and turning whatever survives into something that holds up under real traffic.
The pool is genuinely small and almost entirely visible: they publish, they contribute to open source, they post training runs. Which means every well-funded lab is already writing to them, usually with the same message.
So the constraint is not discovery. It is being worth replying to, which comes down to whether your problem is interesting and whether the message proves you understood their work.
How the agent runs a research engineer search
It reads the work, not the citation count
Papers, yes, but also the repository behind them, benchmark contributions, and write-ups on what did not work. Reproducibility is the signal.
It knows who has already been asked
The most-cited names get dozens of identical messages. The agent also surfaces the second author who did the engineering, who is often the better hire and rarely contacted.
Outreach that engages with their research
Each message references the specific result and connects it to the problem you are working on. Anything less does not get opened twice.
Screening on the translation half
Your questions go out early: a result they failed to reproduce, and the last research idea they killed because it would not survive production.
Questions.
Answered.
Not necessarily. Plenty of excellent research engineers have never published and are visible through open-source work and reproduction efforts instead.
Usually on the problem and on ownership rather than on cash. The outreach leads with what they would work on and what they would decide, because that is what actually moves people.
No. A research scientist is measured on novel results; a research engineer is measured on making results real. Tell us which and the agent weights accordingly.
Longer than most. The pool is small and heavily courted, so the win comes from writing to the right fifty people properly rather than five hundred generically.
Hire a Machine Learning Engineer
Everyone claims the title. The gap between training a model and running one is where the hire is won or lost.
Hire an AI/ML Engineer
The applied ML talent who turns research into product.
Superposition for AI Startups
Hire builders who ship AI products, not resume keyword-stuffers.
Find your research engineer.
Tell us the problem you are working on and the agent will find the people already close to it.