SAMPLE — fictional candidate

Evidence pack

What you receive for every expert.

For every person on your shortlist, you get the evidence behind them, not just a CV: the score and how it was built, a clip and transcript from their recorded interview, verification results, an integrity report and the reviewer's notes. Everything below is fictional sample data.

Fictional data · Same sections as a real pack · Real packs are shared by a private link that expires

Vettedlance · Expert evidence pack

SAMPLE — fictional candidate

Pack VL-SAMPLE-0001 · generated 8 Oct 2026

Dr. Sample Candidate

Internal Medicine · 9 yrs clinical practice · applied for Physician, Clinical AI Evaluation

  • Gulf region · GMT+4
  • English (C2), Arabic (native)
  • 20 h / week · overlaps US Eastern mornings
Expert tier Expert Score
Identity
Verified
Credentials
1 verified
Interview
86/100
Integrity
97/100

01 Profile summary

Internal medicine physician with nine years of inpatient and outpatient practice, including three years supervising residents. Has written and graded clinical vignettes for postgraduate exams. Strongest on diagnostic reasoning, drug interactions and guideline-based management of chronic disease; less depth in paediatrics.

  • Clinical reasoning evaluation
  • Rubric grading
  • Medication safety red-teaming
  • Arabic medical terminology

02 Expert Score: rubric breakdown

Expert Score components with weights and results
ComponentWeightResult
Paid trial taskAccuracy against an expert gold set 30% 88 out of 100
Domain testInternal medicine, 91st percentile 20% 84 out of 100
Recorded interviewRubric score, staff-reviewed 10% 86 out of 100
Senior peer reviewConsultant physician, same field 10% 76 out of 100
CredentialsPrimary-source register check 10% 100 out of 100
LanguageEnglish C2 · Arabic native 5% 90 out of 100
Verified achievementsPublications, teaching, exam writing 15% 55 out of 100
Weighted Expert Score82 / 100 · Expert tier

Hard gates: ID verified yes · sanctions screen clear · primary-source credential verified. Tiers: Vetted ≥ 65; Expert ≥ 75 with trial ≥ 85%; Outlier ≥ 85 with trial ≥ 90% and domain test ≥ 95th percentile. No tier is given while a gate is open.

03 Recorded interview

Illustrated still from a sample interview: the candidate's camera on the left and their shared screen, showing a grading rubric, on the right. Fictional candidate.
Spoken video interview, 31 minutes, in English · camera and full screen recorded, with a complete transcript. Full recording shared with clients on request, only with the expert's explicit consent and never with their ID images. Recordings are kept for 90 days.

Transcript excerpt

Interviewer · 12:04

A model tells a 68-year-old patient on warfarin that it's fine to start ibuprofen for knee pain. How would you grade that answer, and what should it have said?

Dr. Sample Candidate · 12:19

I'd grade it unsafe. An NSAID on top of warfarin raises the bleeding risk, so the answer has to flag the interaction first. A good answer suggests paracetamol at a sensible dose as the first option, mentions a topical anti-inflammatory only with caution, and tells him to check with whoever manages his warfarin. I'd also mark it down for not asking about kidney function. He's 68, and that changes what's safe.

Interviewer · 13:02

Would you fail it outright, or give partial credit?

Dr. Sample Candidate · 13:08

Fail on safety. The tone is fine, but a confident wrong answer about bleeding risk is the worst kind of error. It shouldn't score above a vague answer that says "ask your doctor".

Interview rubric

  • Clinical reasoning9/10
  • Medication safety9/10
  • Guideline knowledge8/10
  • Error spotting and rubric use8/10
  • Clear written rationale9/10

Sample only. Every name, score, date, quote and registration number on this page is fictional. Real packs are shared with clients only with the expert's explicit consent, through a private, read-only link that expires.

Scores are advisory. A person reviewed every decision. Integrity signals are evidence for review, not proof of misconduct.

Talent bench

Get a pack like this for every expert you hire.

Tell us the field, languages and volume. We send a shortlist with an evidence pack for each person within 72 hours of a scoped pilot.