About

Doverank is a people research platform for recruiters. Instead of searching a profile database, an agent researches candidates in the primary sources where their real work lives, then scores everyone against the bar you set.

Why we built it

Recruiting tools got very good at returning lists and never got good at telling you anything. You type filters into a box, you get four hundred profiles back, and then you do the actual work by hand: opening tabs, reading between the lines of a job title, trying to figure out whether the person who calls themselves a machine learning engineer has ever shipped a model.

The information that answers those questions is usually public. It is in the papers someone published, the code they wrote, the trials they ran, the filings their fund made. It is just scattered across a dozen places no one has time to check on a Tuesday afternoon.

So we built the thing we wanted: an agent that goes and reads all of it, keeps what holds up, and hands back a short list with the reasoning attached. You still make the call. You just get to make it knowing something.

What is Doverank?

Doverank is a research agent for recruiters, wrapped in a table you can work in. You describe who you are looking for in plain language, paste a job description, or drop the profile of someone who already fits. The agent turns that into a screening checklist you can edit, searches the sources that matter for the role, verifies what it finds, and fills a live table with candidates. Every row carries the evidence behind it, a match score against your checklist, and a signal about whether the person looks open to moving.

How is Doverank different from a search tool?

A search tool takes a query and returns rows from a database. Doverank runs a loop. It qualifies companies against real funding data before it looks for people at them. It resolves identities before spending anything on enrichment, so you do not end up emailing the wrong person with the same name. It reads the work itself rather than trusting a job title. And when a search comes up empty, it says so instead of padding the table with people who nearly fit.

What Doverank is not

It is not an applicant tracking system, and it does not want to replace yours. It is not a mass outreach tool, and it will not send a thousand sequenced emails on your behalf. It does not scrape private data or anything behind a login. It reads what is publicly available about someone professionally, which is the same material a thorough human researcher would read, and it does that in minutes instead of days.

What data sources does Doverank use?

The agent picks its sources based on the role you are hiring for. A physician search runs through federal registries. A research search runs through publication records. These are the places it reads.

Semantic Scholar

Papers, citations, and co-authors across every field of research.

OpenReview

Submissions and peer reviews from the major machine learning conferences.

arXiv

Preprints, often months before the work appears anywhere else.

GitHub

Repositories, contribution history, and the code behind a claim.

Hugging Face

Models and datasets people have published, and how widely they are used.

Kaggle

Competition results and rankings for data science roles.

NPI Registry

The federal registry of licensed clinicians in the United States.

ClinicalTrials.gov

Who leads which trials, and who is listed as a sub investigator.

PubMed

Biomedical publication records and author affiliations.

SEC EDGAR

Filings that reveal the people behind funds and growth stage companies.

Y Combinator

The startup directory, including who is hiring right now.

CourtListener

Federal dockets showing which attorneys actually litigate what.

FAA Registry

Certificated pilots in the United States.

LinkedIn

Career history, used to connect the rest of the evidence to a person.

Who is Doverank for?

Agency recruiters

Search firms that live or die on the quality of a shortlist, and need to show a client why each name is on it.

In house talent teams

Small teams hiring for roles where the good candidates are not applying to anything.

Technical recruiting

Machine learning, research, and engineering roles where the work sits in papers and repositories rather than on a profile.

Healthcare and life science recruiting

Physicians, principal investigators, and scientists who are documented in public registries and trial records.

Executive and specialist search

Roles where a handful of people in the country qualify, and finding all of them matters more than finding many.

Know who you're calling before you call