Why blind hiring matters
Traditional resume review exposes recruiters to names, photos and prestige signals before any skill evaluation. Those cues disproportionately harm candidates from non-traditional backgrounds. RecruitAI’s approach is architectural: blind mode runs as a pre-processing guardrail inside the ingestion pipeline. Redacted content is what the evaluator agent sees; the original is retained (RLS-isolated) only for post-screen reveal and compliance audit.
- Full name, photo, contact identifiers
- Location, graduation year, institution prestige
- Gendered pronouns and other demographic proxies
- Architecture patterns & system design depth
- Language/tooling proficiency with evidence
- Measured impact: scale, latency, revenue, users
Every redaction and score is logged with timestamps. Export hiring dossiers with bias-audit disclosures and EEOC non-discrimination notices.
Toggle per requisition — reveal when ready
Blind mode is not global: set it per campaign. Enable for top-of-funnel screening to maximize fairness, then reveal candidate identities when you shortlist for interview. Recruiters remain in control, and the system never re-exposes PII to the evaluation agent after reveal.
FAQ — Blind hiring
What exactly does blind mode redact?▾
Names, email addresses, phone numbers, photos, locations, graduation years, university prestige markers and gendered pronouns. Only technical accomplishments, stack depth and project impact remain for scoring.
Can recruiters reveal identity after scoring?▾
Yes. Identity is hidden during evaluation and revealed only when you move a candidate to interview. This preserves auditability while restoring context for final conversations.
Does blind hiring affect accuracy?▾
It increases accuracy for role-relevant skills. Models evaluate architecture, code impact and tooling depth — not demographic proxies that should never influence engineering hiring.
Is blind screening compliant with NYC LL144?▾
RecruitAI logs selection and scoring distributions and can generate impact-ratio reports. Combine with human confirmation checkpoints for Article 22 and EEOC-aligned workflows.