Candor
Open roles Pricing Open live demo →
Glass Box scoring Every rank, fully sourced

AI shortlisting
that shows its work.

800 applicants for one support role. Candor ranks every one and shows you exactly why — the evidence behind each score, the guardrails that capped it, and the fairness checks that ran. No black box.

Explore the demo → Browse open roles
candor.hr/jobs/support-specialist/shortlist
PN
Priya Nandakumar
Support Specialist · ranked #1 of 823
84
FIT
CONTRIBUTION LEDGER
● Tooling +22 ● Comms +25 ● Experience +18
Knockout checkpass
Evidence grounded12/12
Fairness tripwireclear
PII stripped before scoring · every claim traces to a resume line.
Auditable, line by line
Each of five weighted criteria shows its exact contribution to the score. Nothing is hidden in an embedding.
Grounded in evidence
Every claim links to a verbatim resume quote. A claim that can't be traced is flagged, never surfaced as fact.
Fair by construction
Deterministic guardrails cap over-confident verdicts, a bias tripwire forces human review, and adverse-impact is monitored per role.
THE COST-TIERED FUNNEL — MOST CANDIDATES NEVER TOUCH AN LLM
823
Applied
parse · PII strip
→
318
Hard filters
free · deterministic
→
64
Ranked
semantic match
→
64
Agent panel
Skills ‖ Risk ‖ People
→
9
Shortlist
est. spend $3.90
Open-source ATS · transparent, bias-audited AI shortlisting
Candor
Recruiter sign in →

Open roles at Northwind

Every application is scored transparently. You'll always know where you stand.

{{ j.title }} {{ j.type }}
{{ j.location }}{{ j.salary }}{{ j.applicants }} applicants
← All roles

Customer Support Specialist

Remote · Global$46k–$58kFull-time

You'll be the first human our customers reach when something breaks. Tier-2 troubleshooting across our SaaS product — email, chat, and the occasional call — owning tickets end to end and turning frustrated users into fans.

What we're looking for
  • 2+ years in a customer-facing support role
  • Hands-on with a helpdesk platform (Zendesk, Intercom, or similar) — must-have
  • Clear, warm written English (C1+)
  • Comfortable troubleshooting technical issues methodically
  • Bonus: a second language for regional coverage
Application received
Your resume is being scored right now against five transparent criteria. You'll see your standing — and the reasoning — within the hour.
← Back to roles
Apply for this role
Drag a PDF or DOCX here — we strip contact details before any AI reads it
Your name, email and phone are never shown to the scoring model.
Candor
Northwind Labs
{{ n.icon }}{{ n.label }}
ACTIVE ROLES
{{ j.title }}
RC
Rina Castillo
Recruiter
{{ crumb }}
Search candidates…
{{ themeIcon }}

Good morning, Rina

2 active roles · 1,204 applicants in the pipeline · 3 flagged for your review

{{ s.label }}
{{ s.value }}{{ s.delta }}
Pipeline — Customer Support Specialist
823 applied · funnel to shortlist
{{ f.label }}{{ f.count }}
Adverse-impact monitor
Four-fifths rule · voluntary self-ID · never used in scoring
{{ a.group }}{{ a.rate }} ({{ a.ratio }})
{{ a.flag }}
Flagged for human reviewOpen shortlist →
{{ r.initials }}
{{ r.name }}
{{ r.reason }}
{{ r.fit }}
{{ r.chip }}

Jobs

2 open · 1 draft

New job
{{ j.title }}
{{ j.meta }}
{{ j.applicants }}
applicants
{{ j.status }}
Jobs /

Customer Support Specialist

Open 823 applicants · closes in 9 days
{{ t.label }}{{ t.badge }}
{{ s.label }}
{{ s.value }}
Scoring rubric
{{ c.label }}MUST-HAVE
{{ c.weight }}
Recent activity
{{ a.text }}
{{ a.time }}
{{ col.label }} {{ col.count }}
{{ c.initials }}
{{ c.name }}
{{ c.region }}
{{ c.fit }} {{ c.chip }}
+{{ col.more }} more
Ranked shortlist64 scored
{{ r.rank }}
{{ r.initials }}
{{ r.name }}
{{ r.sub }}
{{ r.fit }}
{{ cand.initials }}
{{ cand.name }} PII stripped
{{ cand.sub }} · ranked #{{ cand.rank }} of 64 scored · applied {{ cand.applied }}
{{ cand.baseFit }}FIT
RECOMMENDATION
{{ cand.recIcon }}{{ cand.recLabel }}
▼ capped from “{{ cand.cappedFrom }}”
{{ cand.reviewIcon }}
{{ cand.reviewTitle }}
• {{ rr }}
Contribution ledgerhow the {{ cand.baseFit }} was assembled
weighted mean · 5 criteria
+{{ seg.points }}
{{ seg.label }} {{ seg.math }}
{{ cand.ledgerNote }}
Verdict reconciliation
The model proposed a verdict. Deterministic guardrails reconciled it against hard facts — caps only, never upgrades.
{{ step.label }}
{{ step.note }}
→
EVALUATOR AGENTS — RAN IN PARALLEL, EVIDENCE-CITED
{{ ag.icon }}
{{ ag.name }}
{{ ag.model }}
{{ ag.score }}/100
▼
{{ ag.summary }}
{{ f.tagIcon }}
{{ f.claim }}
{{ f.evLabel }}
w{{ f.weight }}
Resume — grounding source {{ cand.groundedCount }}
{{ ln.text }}
{{ cand.evidenceHintText }}
Deterministic guardrails
toggle to see the black-box verdict
{{ g.icon }}
{{ g.label }}
{{ g.detail }}
Fairness tripwire{{ cand.fairStatus }}
{{ cand.fairIcon }}
{{ cand.fairText }}
Flagged phrase: “{{ cand.fairTerm }}” in the synthesis rationale. Score held; routed to you.
Integrity checks
{{ i.icon }}
{{ i.label }}
{{ i.detail }}
Your decision
{{ cand.decisionNote }}
{{ toast.msg }}