AI System Transparency
Last updated: 2026-03-08
1. System Purpose
Semantic search and ranking of candidate CVs/resumes to assist human recruiters in identifying relevant candidates.
2. Risk Classification
Under the EU AI Act, AI systems used in recruitment and employment decisions are classified as high-risk (Annex III, paragraph 4). RecruitSecure AI is designed and operated in compliance with the requirements for high-risk AI systems.
3. AI Model
| Model | multilingual-e5-small |
| Dimensions | 384 |
| Languages | 100+ (multilingual) |
| Quantization | q8 (8-bit) |
| Processing | 100% local — no external API calls |
4. Data Processing
- Input types: PDF, DOCX, TXT
- Storage: EU (Supabase, Frankfurt region)
- External API calls: None — all AI processing happens on our servers
- Encryption: AES-256-GCM at rest
5. Human Oversight
All AI-generated rankings are advisory only. Human recruiters make all hiring decisions. The system does not autonomously accept or reject candidates.
6. Limitations
- Ranking is based on text similarity, not candidate quality assessment
- Results may vary by language — model trained on multilingual data but not equally across all 100+ languages
- Maximum 512 tokens per document passage are considered for embedding
- System does not evaluate soft skills, cultural fit, or non-textual qualifications
- Duplicate or very similar CVs may receive near-identical scores
7. How Rankings Work
- Your search query is converted into a 384-dimensional semantic vector using the multilingual-e5-small model.
- Each candidate CV has been previously converted into the same vector space during upload.
- Results are ranked by a hybrid score combining cosine vector similarity (semantic meaning) and BM25 full-text matching (keyword overlap).
- For Pro and Enterprise plans, ML-powered reranking may adjust scores based on your organization's historical feedback (thumbs up/down on previous results).
- All rankings are advisory — the system does not make hiring decisions.
8. Your Rights
Under the EU AI Act and GDPR, you have the right to:
- Request an explanation of how a specific ranking decision was made
- Request human review of any AI-assisted decision
- Export or delete your data at any time (see Privacy Policy)
- Lodge a complaint with your national supervisory authority