Stop debating what a score means
Abstract numbers don't drive hiring decisions. SkillFoundry benchmarks every candidate against your own team — overall, by dimension, and on AI orchestration. Scroll down to see the same panel reviewers see.
How it works
Your bar, applied to every candidate
Benchmarking reframes every result: not 'they scored 84' but 'they exceed most of your backend engineers on this task.'
Calibrate your cohort
Your engineers complete the same assessment instrument candidates take. Their scores form the reference cohort — with consent.
Score every candidate
Each submission gets an overall score plus dimension and AI-orchestration sub-scores — the same rubric for everyone.
Read percentiles, not abstractions
"82nd percentile vs. your backend team" is actionable. A bare 84/100 is not — unless you know what your team typically scores.
What you can read
Every percentile tells a story
Hiring committees use these comparisons to decide whether someone clears the standard your best engineers set.
Overall percentile
Where this candidate sits against your calibrated engineering team on the composite score.
Example: 82nd percentile — stronger than most of your backend engineers on this task
Dimension breakdown
See whether they are strong on code quality, correctness, tests, or behavioral signals — not just the headline number.
Example: Verification discipline: 88th percentile
AI orchestration vs. team
Compare delegation, verification, recovery, and judgment against engineers who already work with AI on your stack.
Example: Exceeds 9 of 12 backend engineers on AI orchestration
Fairness guardrails
Comparisons are withheld when the cohort is too small or an adverse-impact check flags the population.
Example: Frontend cohort (5 engineers) — comparison unavailable until 8+ consented members
Demo mode
This page uses canned benchmark data. The panels below are the same CohortBenchmarkPanel reviewers see on evaluated submissions — nothing is simulated on the frontend beyond the sample dataset.
Panel 1
Healthy benchmark cohort
Alex Chen scored 84/100 on a pagination task. Against your 12-person backend cohort, that translates to an 82nd-percentile overall — and exceeds 9 of 12 on AI orchestration.
Team benchmark
vs your team (12 engineers)
Overall
82nd
Score dimensions
AI orchestration
Decision-support only — percentiles compare this candidate to your calibrated engineering team, not a pass/fail bar.
Panel 2
When comparisons are withheld
SkillFoundry fails closed. Small cohorts and adverse-impact flags suppress percentile views rather than showing misleading comparisons.
Team benchmark
Benchmark unavailable — need at least 8 consented engineers in "Frontend team" (currently 5).
Team benchmark
Benchmark comparison withheld: adverse impact detected for Female (sex) (four-fifths rule). Consult your compliance team before using percentile comparisons.
Ready to benchmark candidates against your own engineers?
Calibration takes one assessment per reference engineer. After that, every candidate result includes percentile context.