Describe the candidate Let an AI agent screen 800M+ public profiles
Move from a vague job description to high-caliber, under-the-radar candidates in minutes thanks to AI agents combined with Nova's proprietary merit-based search engine
You don't have to remember everything upfront, or stay on one tab to fix it.
What recruiters run into every day
You're scrolling through hundreds of profiles that don't matter.
Most searches return a wall of loosely-relevant names, and the only way to find the handful worth reaching out to is opening every single one.
Search "Banking," and a regional branch and a Global Systemically Important Bank rank the same.
A great candidate never shows up, because they didn't use your exact words.
Keyword search only finds people who described themselves the way you searched for them. Anyone who phrased it differently just isn't in the results, qualified or not.
Search "ML Engineer," miss the "Applied Scientist" who's just as qualified.
Every good candidate lives on a different platform, and you're checking each one by hand.
LinkedIn, GitHub, niche job boards. The person you need could be on any of them, so sourcing turns into ten open tabs and a spreadsheet just to keep track.
How the agent learns what you want
Start from any search.
The agent proposes Mandatory and Desirable criteria straight from your filters. Drag them into the order that matters to you.
Calibrate it.
It shows you real candidates from your search, one at a time, "Would you consider this a valid candidate?" Say yes or no. If something's off, tell it right there ("I want to exclude fashion") and it adjusts before it's seen the rest.
Set the scope.
Pick the maximum number of candidates you want, and whether they land in a new list or an existing one. If fewer candidates qualify, that's fine. The agent won't pad the list just to hit a number.
Put the agent to work.
It runs through the rest of your search on its own, applying everything you just calibrated, and saves only genuine matches.
The impact
- 2x
- Higher-quality matches
- +12–15h
- Saved per week on search work
Every candidate is screened against your exact criteria before they reach your list, not just keyword-matched.
Hours that used to go into opening profile after profile, back in your week for the calls that actually close a placement.
Two scores, one clear view
Nova's own quality signal, built on 150,000+ assessments from its recruiter network. It weighs international experience, languages, education, and career trajectory, independent of any one role, so a strong candidate reads as strong no matter what you're searching for.
How well this candidate fits this search, specifically. Recalculated every time.