Scoring Engine
How Vetriva evaluates candidates and produces a scored recommendation.
How scoring works
Every candidate in Vetriva has two scores: a baseline score derived from AI resume analysis, and an adjusted score that incorporates your team's feedback signals. The adjusted score drives the final recommendation.
Baseline score
When you upload a resume, Vetriva's AI extracts structured features — skills, experience, education, seniority signals — and compares them against the role context. The result is a number between 0 and 100 representing initial fit.
The baseline score does not change after it is assigned. It is a fixed reference point.
Signal adjustment
Each feedback signal carries a delta — a positive or negative value that modifies the score. The delta depends on both the signal value and the confidence level:
| Signal | Base delta |
|---|---|
| Strong Yes | +2 |
| Yes | +1 |
| Neutral | 0 |
| No | −1 |
| Strong No | −2 |
The base delta is multiplied by the confidence level (1–5), then scaled and clamped so the adjusted score stays within 0–100.
Decision thresholds
The adjusted score maps to a recommendation according to fixed thresholds:
| Adjusted score | Recommendation |
|---|---|
| 80 – 100 | Strong Consider |
| 60 – 79 | Consider |
| 40 – 59 | Borderline |
| 0 – 39 | Reject |
Example
A candidate has a baseline score of 72 (Consider). Your team adds two signals:
- Strong Yes, confidence 3 → +6 points
- No, confidence 2 → −2 points
Net adjustment: +4. Adjusted score: 76 → still Consider, but higher confidence. A third signal of Strong No at confidence 5 (−10) would push the score to 66 — still Consider but approaching borderline.
Score freshness
Scores are recalculated in real time as new signals are added. There is no manual refresh needed. Older signals carry the same weight unless explicitly overridden by a newer signal from the same reviewer.