How PATHS Works
Fair hiring, step by step
Every phase is designed to reduce bias, increase transparency, and keep humans in control — from first CV upload to final hire.
01
Create your profile
— Upload once. We do the rest.- Upload your CV (PDF or DOCX) — AI extracts your full work history, skills, and education
- Add your LinkedIn, GitHub, and portfolio links for richer evidence
- Complete short onboarding steps to confirm and enrich extracted data
- Set your job preferences: role type, location, salary range, remote/hybrid
- Your profile is saved and reusable across multiple job opportunities
02
AI extracts & organizes
— Evidence over inference.- CV parsing agent extracts every skill, role, achievement, and credential
- Each extracted fact becomes an 'evidence item' — traceable to source text
- Skills are normalized and categorized (technical, soft, tools, languages)
- GitHub and portfolio links are analyzed for additional technical evidence
- All evidence is stored with confidence scores — nothing is fabricated
03
Smart job matching
— 15 agents, one transparent score.- 15 specialized AI agents score your profile against each job's requirements
- Skills match, experience fit, education level, and culture signals are all weighted
- Every score is backed by evidence — you see exactly which skills earned which points
- Vector similarity finds candidates who match conceptually, not just by keyword
- Scores are deterministic and auditable — no black-box decisions
04
Fair, anonymous screening
— Your skills speak. Not your name.- Recruiters see skills, experience, and scores — not name, photo, or personal details
- Anonymized view hides: full name, email, phone, exact address, nationality
- Bias flags automatically detect and surface potential discrimination risks
- De-anonymization requires explicit recruiter request + HITL approval
- Full audit trail maintained for every view, decision, and de-anon request
05
Human makes the final call
— AI proposes. Humans decide.- Every shortlist is reviewed by a human recruiter before candidates are contacted
- Human-in-the-loop approvals required for outreach, shortlisting, and final offers
- Recruiters see AI rationale alongside their own judgment
- Candidates are never automatically rejected — a human always reviews edge cases
- Decision packets include evidence, scores, and bias check results
Built on six principles
Transparent
Every score is backed by evidence. No opaque AI decisions.
Private
Candidate PII is protected by anonymization until consent.
Compliant
Egypt PDPL, EU AI Act, and EEOC guidelines enforced by design.
Fast
From CV upload to ranked shortlist in minutes, not weeks.
Auditable
Full audit log of every hiring action — exportable on demand.
Human-first
AI assists. Humans own every decision that affects people.