🎯BurntCV Hire// SHOWS ITS WORK
Open workspace →
AI Recruiting

Hiring becomes faster with AI that shows its work.

BurntCV Hire screens résumés, identifies top candidates, and matches applicants against the role's real bar — with per-requirement scores, quoted evidence, and interview questions aimed at the gaps. The AI analyzes. You decide.

Screen your first candidate — freeOur data promise
Free pilot: 1 role · up to 5 candidates · full fit reports. No card.
Kubernetes production experience MUST-HAVE · WEIGHT 9
2 / 4 · Partial
✅
“Deployed microservices on Docker Swarm serving 2M requests/day” — adjacent container-orchestration evidence, requirement not directly met.
⚠️
Lists “Kubernetes” in skills — no production usage found anywhere in work history.
Ask: “You list Kubernetes but I don't see production usage — walk me through the last cluster you operated and what broke.”
↑ an actual dimension from a fit report — evidence, contradiction, and the question to ask.

Decision support, not a black-box gatekeeper.

01
Paste the JD — you set the bar
We decompose the job description into weighted requirements: must-haves, nice-to-haves, implicit signals. You review, re-weight, and confirm every one. The AI proposes; the recruiter defines what matters.
02
Add candidates
Paste résumé text or drop a PDF, attest the consent basis, and the pipeline extracts a structured profile — every fact carrying the verbatim line it came from.
03
Read the evidence. You decide.
Each requirement is scored 0–4 with quoted supporting and contradicting evidence, gaps surfaced, and interview questions targeted at exactly what's unproven. Advance, hold, or pass — logged, with your name on it.
🧾
Every score cites its line
No holistic vibe-scores. Per-requirement ratings, each backed by a verbatim quote from the résumé — supporting and contradicting.
🧮
The verdict is math, not mood
The AI extracts evidence; deterministic code applies your weights. Same input, same score, every time — reproducible and defensible in an audit.
🙋
A human owns every decision
No silent auto-reject. Ever. Knockout failures are flagged, never actioned — every advance, hold, and pass is recorded to a named person.
🛡️
DPDP-first by design
Consent attestation per candidate, purpose-limited to this role's screening, retention windows with real deletion, and one-click export for access requests.
Two products. One hard wall between them.
The roast never stores a résumé — that promise stands. Hire stores candidate data for your role only, in your tenant, deleted on your schedule. Nothing ever crosses between the two. Not a résumé, not a pipeline, not a table.
Start screening →