Company
Hiring did not get harder. It got louder.
Applications per job doubled while recruiting teams halved. Neuradesk Hire is built for the part of that problem software can actually solve: telling signal from volume, and proving where the signal came from.
Why we built it
The volume side of the same story
Neuradesk started on the candidate side, because a founder sent more than a hundred applications and got three or four replies. Working out whether that was normal turned up the other half of the picture, the half recruiters live in.
Between 2022 and 2025, applications per job on one major hiring platform rose from 116 to 244. Over the same window the number of recruiters per company fell from about ten to under five, and time to fill went from roughly 44 days to nearly 60.
Nobody on either side of that is doing a bad job. There are simply twice as many applications arriving and half as many people left to read them.
Twice the volume, half the team. Almost every hiring problem downstream of that is a reading problem.
What we found
The screening story you have been sold is the wrong one
The most repeated statistic in this industry is that applicant tracking systems reject 75 percent of resumes before a human sees them. There is no study behind it. It traces to a 2012 sales pitch by a resume-optimisation vendor that shut down the following year, and the citation trail behind it ends in nothing.
Read what the systems actually do and the picture changes. The major vendors' own documentation describes automatic rejection as something triggered by a candidate's answers to application questions, never by resume text.
Screening software mostly does not reject candidates. It fails to surface them.
Which means the filter that really decides your pipeline is your own saved search, and the ranking that decides which twenty of two hundred get read. That is a product problem, and it is the one we work on.
What we built
One graph, not ten dashboards
Most recruiting tools sell an isolated artefact: a search, a scorecard, an inbox. Ours share one candidate graph, so a search knows what your last five scorecards actually rewarded, and an interview knows what the resume claimed.
The reason that matters is not tidiness. Once AI made application text cheap, one study found the link between polished wording and callbacks fell by half, and employers shifted toward verifiable work history instead. A profile a candidate types is marketing. A profile confirmed at the source is evidence.
What that looks like in practice
Candidates arrive with a profile that was assembled and checked on the other side of the same system. Every recruiter action writes to the same audit trail as it happens. Nothing has to be stitched together out of exports at the end of a quarter.
What we will not do
Things we decided against
We do not publish customer counts we cannot substantiate.
No logo wall of companies that trialled us once, no invented case studies, no stock-photo testimonials. Neuradesk Hire is founder-led and in private beta, and we would rather say that than dress it up. When we have numbers worth showing, they will carry a method.
We do not sell candidate data, and we never will.
There is no data broker, no ad network and no retargeting pixel anywhere in the product. Candidates you source stay yours, and stay theirs.
We do not claim candidate data never leaves.
AI features send resume text to third-party model providers. We name every one of them on the sub-processor page, with the terms each of them publishes, rather than implying nobody sees it.
The founders
Who is building this
Nischith Kashyap S
Founder and CEO
Built the first version of Neuradesk after his own search stalled at a hundred applications. Runs product.
Sriram S
Co-founder and CFO
Chartered Accountant. Owns the numbers, the contracts, and everything a buyer will eventually ask to see.
Neuradesk AI Pvt Ltd, incorporated 2026, Karnataka, India.
Trust
What happens to candidate data
- Every organisation's data is isolated at the database layer.
- Tenant queries run under a restricted Postgres role with row level security forced on, not under an application filter somebody can forget to write.
- Every resume a recruiter opens writes an audit row.
- The row is written inside the same transaction as the read, then signed and chained to the entry before it, so tampering is detectable.
- Candidates are private until they turn discovery on.
- Search never returns a private profile, and we record which version of the consent notice each candidate agreed to.
- AI features send resume text to model providers, and we name all of them.
- We hold no zero-retention agreement with any of them, so we publish the list and each provider's own terms instead of implying nobody sees it.
- Your card details never touch our servers.
- Payment goes into Stripe's and Razorpay's own hosted forms. We store no card number, no last four digits, no expiry.
- Transport and sign-in are locked down by default.
- One year of HSTS with an enforcing content security policy, password hashing at bcrypt cost 12, and passkeys on every tier.
- jobs indexed
- 19,000+
- languages
- 14
- pricing regions
- 11
- building since
- 2024
FAQ
Questions buyers ask first
What is Neuradesk Hire?
Neuradesk Hire is the recruiter side of Neuradesk.ai, built by Neuradesk AI Pvt Ltd in Karnataka, India. It combines candidate search, screening, interviews, a shared inbox and a signed audit trail on one candidate graph, so that each part informs the others.
Is Neuradesk Hire available now?
Neuradesk Hire is founder-led and in private beta. We onboard a small number of teams at a time so we can build alongside real recruiters, and every plan starts with a 14-day trial that does not ask for a card.
Who can see a candidate's profile on Neuradesk?
Candidate profiles are private by default. A recruiter cannot discover a candidate until that candidate turns discovery on, and Neuradesk records which version of the consent notice they agreed to.
What happens when a recruiter opens a resume?
Every resume view writes an audit row in the same database transaction as the read. The audit log is HMAC signed and each entry is hash-chained to the one before it, so tampering is detectable rather than merely discouraged.
Do applicant tracking systems really reject 75 percent of resumes?
No. That figure traces to a 2012 sales pitch by a resume-optimisation vendor that closed in 2013, and no study supports it. The major vendors document automatic rejection as being triggered by a candidate's answers to application questions rather than by resume text. The realistic risk is that a good candidate never appears in a recruiter's filtered search.
Where is Neuradesk based?
Neuradesk AI Pvt Ltd is registered in Karnataka, India, and was incorporated in April 2026. It is recognised as a startup by the Government of India's Department for Promotion of Industry and Internal Trade (DPIIT), certificate DIPP277125. Work on the product began in 2024.
See what your pipeline actually looks like in 14 days.
Start a trial without a card. Bring one open role and we will help you set it up on a call.