AI-native hiring decision system

Know who to interview.Know why.

Cernor helps your team agree on what good looks like, reads every candidate against the same bar, shows what the evidence actually proves and follows up where something important is still unknown.

Start free See how it works

One role. Up to five candidates. Real analysis.

Cernor holds one open role as thirteen unsorted signals at once — quoted language from the brief, system and compliance terms, and concepts, all on screen together: “a regulated environment”, “a People Ops function from zero”, “every clinician we bring on”, “no team to inherit”, “spreadsheets and Slack”, exec headcount planning, PHI in every system, exec staffing reviews, multi-state employment, leave administration, people-data audit trail, clinician onboarding, and immigration and visas. The mass gathers from its own centre outward, and the whole of it clears once the reading finishes, so the conclusion resolves into the space it leaves. Below it the record stack waits, unread and unnamed. A reading pass crosses the field, and the three quoted fragments that carry the decision — “a regulated environment”, “a People Ops function from zero” and “every clinician we bring on” — resolve to ink and keep an annotation rule. From those, the role itself is inferred: People Ops Generalist at a healthtech company in San Francisco. Their conclusion is then written once — a zero-to-one People operator with executive range and the judgment to own regulated data — which becomes three criteria. Zero-to-one people ops and exec range with hands-on lock as gates provable from the record; PHI-grade data care stays open and can only be settled in follow-up. The Standard locks, the cohort comes into focus, and 186 candidate records are read against it. Three candidates settle with different decision postures. Alina Serrano scores 84.2 and is Interview Ready: ownership and executive range are supported, data care remains unresolved. Naomi Castellanos reaches 61.8, then moves to 51.4 — a minus 10.4 trust adjustment — and is marked Human Verify, because a claimed scope exceeds the supporting chronology. Devin Park scores 57.3 and is marked Follow-Up: two criteria remain unproven. 186 records read, three candidates on the shortlist, four candidate-specific follow-ups ready.
People Ops Generalist
Healthtech · San Francisco, US
Role signal· inferring
PHI in every system clinician onboarding exec staffing reviews immigration and visas multi-state employment “a regulated environment” leave administration “a People Ops function from zero” “no team to inherit” exec headcount planning “every clinician we bring on” people-data audit trail “spreadsheets and Slack”
What the role actually demands
A zero-to-one People operator with executive range and the judgment to own regulated data.
Zero-to-one people ops
Exec range + hands-on
PHI-grade data care
C-317
Devin Park
Trust —
C-084
Rachel Imani
Trust —
C-129
Todd Bergstrom
Trust —
C-203
Sofia Duarte
Trust —
C-046
Alina Serrano
Trust —
C-268
Aaron Weiss
Trust —
C-355
Leah Nakamura
Trust —
C-172
Naomi Castellanos
Trust —
C-411
Chris Villanueva
Trust —
C-490
Bianca Ruiz
Trust —
C-538
Owen Tran
Trust —
Reading cohort 001 / 186
186
Records read
3
On shortlist
4
Follow-ups ready
People Ops Generalist
Healthtech · San Francisco, US

“A zero-to-one People operator with executive range and the judgment to own regulated data.”

Zero-to-one people ops
Exec range + hands-on
PHI-grade data care
Alina Serrano
84.2
Interview Ready
Devin Park
57.3
Follow-Up
Naomi Castellanos
−10.4 51.4
Human Verify
186
Records read
3
On shortlist
4
Follow-ups ready

What this shows: from unsorted signal — quoted language from the brief, system and compliance constraints, and the record itself — Cernor infers what the role actually requires, then reads the whole candidate cohort against it.

One hiring bar

Agree on what good looks like before anyone gets scored.

Cernor turns the role into a Standard your team can shape, approve and lock before candidate evaluation begins. Every candidate is then judged against that same version.

The bar stops moving from person to person.

One job description for an Operations Manager at a lean Hong Kong logistics company is read straight through. Six phrases in it carry the actual decision, and each becomes a written criterion in an approved Standard. Warehouse, third-party logistics and cross-border lanes becomes Operational Breadth, a gate needing follow-up. Five to eight years in logistics, supply chain or e-commerce operations becomes Logistics Operations Tenure, a gate provable on the résumé. English and Cantonese required becomes Language Proficiency, a gate provable on the résumé. Holding vendors accountable through scorecards becomes Vendor and Partner Accountability, a core signal needing follow-up. Dispatch time, inventory accuracy and cost per order becomes Inventory, Returns and Reconciliation, a core signal needing follow-up. Rebuilding the SOP that caused a delay becomes Process Improvement and SOP Buildout, a support signal needing follow-up. The finished Standard, version one, holds six criteria: three gates, three weighted signals, two provable on the résumé and four carried into follow-up. It is breadth-led and gated, and it does not change until a new version is approved.
Job description· inferring 1 of 1

Operations Manager to own daily fulfilment across our warehouse, 3PL, and cross-border lanes. You will be hands-on: troubleshooting a dispatch delay one day, rebuilding the SOP that caused it the next.

5–8 years in logistics, supply chain, or e-commerce operations required. You will hold vendors accountable through scorecards and quarterly reviews, and you will be measured on dispatch time, inventory accuracy, and cost per order.

English and Cantonese required, Mandarin preferred. Lean SMB environment handling higher-value, return-sensitive SKUs, so comfort with ambiguity matters more than headcount.

Operations Manager

Logistics Supplytech · Hong Kong · SMB

Standard · Drafting Standard v1 · Locked
Essential Gates

Operational Breadth

Gate Followup Screen

Logistics Operations Tenure

Gate Resume Screen

Language Proficiency

Gate Resume Screen
Weighted Signals

Vendor & Partner Accountability

Core Followup Screen

Inventory, Returns & Reconciliation

Core Followup Screen

Process Improvement & SOP Buildout

Support Followup Screen

6 criteria · 2 résumé-verifiable · 4 in follow-up

Every candidate

Give every candidate the same serious read.

Cernor reads the career behind each résumé: ownership, scope, outcomes, chronology, contradictions and what actually transfers to this role.

No keyword ranking. No title matching.

The last candidate gets the same depth as the first.

What this shows: each candidate is read against the six criteria of the approved Standard in context — ownership, scope, chronology, outcomes, contradictions and transferability — rather than by keyword overlap.

A candidate read. One sentence at a time, six claims from the record are scored against three registers — Merit, Trust and Proof — each with its own requirement for this role, marked and numbered on its track at seventy, sixty-two and fifty-five. Every claim moves all three up, down, or not at all. The read ends with Merit at seventy-eight and clear of its requirement, Trust at thirty-one and contradicted, Proof at forty-two and unsourced.
Claims analyzed 01 / 06
Merit
0
Trust
0
Proof
0
Evidence & Trust

See what the résumé proves.And what it only claims.

Cernor keeps fit separate from evidence and credibility, so a relevant profile never hides whether an important claim is supported, inferred, unresolved or contradicted.

Polish is not proof.

What this shows: merit, proof and trust are scored separately, one claim at a time. A confident sentence with nothing behind it does not become evidence, and missing proof stays visibly missing.

Follow-Up

Ask only what the résumé couldn’t answer.

When an important question is still open, Cernor writes the next Follow-Up from that exact gap in the evidence.

Your team sees what it will ask and why before anything is sent.

Follow up only where the decision is still open.

An amber point appears alone on the left — the one thing the record could not settle. A line grows out of it and stops in empty space. That stop is measured: a scanner sweeps it, rings pulse, and the point splits into three possible continuations that grow at equal weight. Each is weighed in turn, then two retract into short dead stubs and the third thickens and keeps growing, as the same unbroken stroke, until it stops somewhere new. It happens three times, faster each time, and no part of the route exists on screen before the line arrives at it. At the end a sweep runs the whole length and the colour resolves from violet to green: one determined path. Below, fit moves from seventy-three point six to eighty-one point two, and the next action from Follow-up to Interview.
One thing the record cannot settle The next question does not exist yet Three continuations, none of them wrong The answer closes two and keeps one Deeper, and the shape cannot be guessed
Fit against the Standard 73.6+7.6
Next action Follow-up → Interview

What this shows: the one thing the record could not settle becomes the next question. Nobody wrote it in advance — it did not exist until the evidence ran out at that exact point.

Current decision

New evidence can change who you interview.

As Follow-Up strengthens, weakens or leaves a claim unresolved, Cernor updates the candidate and shows exactly what changed.

The movement stays attached to the evidence that caused it.

The best application is not the best candidate.

Six shortlisted candidates for a clinical trial operations role, graded twice. First on evidence: seven criteria each, taken only from what the record supports. Then on adaptive Follow-Up: a question written from every unresolved point that still decides something, graded on the answer that came back. Fit and decision follow the cells, and the order changes. Callum Reid grades highest on the record at 84.2 and finishes fifth at 61.5, because a delivery metric was contradicted at Follow-Up; he goes to human review. Nadia Okonkwo rises to first at 88.1 and Rowan Ellery to second at 84.6, both to interview. Dev Raghunathan rises from fifth to third at 78.9. Sasha Pemberton returned no answer and is held at 70.2. Imogen Vance is a low match at 52.7 on unevidenced scope. Three go to interview.
Clinical Trial Operations Manager · Toronto Ready
Candidate EvidenceFollow-Up FitDecision
Supported Credible Partial Missing

Decision Packet · Interview focus · What changed

Beyond application review

A rankingis a moment.

The decisionkeeps moving.

Cernor keeps the hiring bar, evidence, open questions, Follow-Ups and decision history together, so new evidence or a changed requirement updates the work without starting over.

See the Cernor system

Approved barSource-linked evidenceVersion historyHuman final call

Put your next role through Cernor.

Bring the role and up to five candidates. See the real analysis on your own hiring problem.

Start free Book a live walkthrough

No ATS integration required.