Hiring platform
Ranking and criteria
Updated September 5, 2026
Every candidate's match percentage comes from ranking criteria set on the job, not from a single opaque model score. Each component is visible in the reasoning trail attached to that candidate.
The four components and their default weights
| Component | Default weight | What it measures |
|---|---|---|
| Skill coverage | 45% | How many of the job's listed skills the candidate's history demonstrates |
| Relevant experience | 30% | Time spent in roles judged adjacent to this one, not just identical titles |
| Trajectory | 15% | Whether the candidate's recent moves trend toward this kind of role |
| Mandatory requirements | 10% | Explicit must-haves; a failure here caps the total rather than subtracting from it |
Where weights are set
Job settings → Ranking criteria. Weights are per job, never per workspace, so a senior and a junior req for the same skill set can weigh experience differently. Changing the weights re-scores the existing batch in place — no re-upload is needed.
A candidate showing an unexpectedly low score with every skill matched has almost always failed a mandatory requirement, which caps rather than averages. Check the reasoning trail's cap marker before assuming the model is wrong.
Re-scoring history
Previous scores aren't overwritten when weights change — they're kept in the audit log, so you can show exactly what the criteria were at the time a candidate was screened out or advanced.
Was this page clear?Report an error in this page