Radiant Revive — Aesthetics & Wellness ← Skin Wellness Home
Equity Audit Service · For AI Developers & Health Systems

Your dermatology AI has a skin-tone story.
Can you prove it's a good one?

Independent fairness and reliability audits for dermatology and skin-imaging AI — built on a published, preregistered methodology, run by the team that discovered why most bias audits can't be trusted.

Book a 30-Minute Scoping Call See Engagement Tiers

🔒 Not ready to talk? Run the free, in-browser Label Reliability Check on your own dataset →

47.9%
How often the field's two expert skin-tone label sources fully agree on the standard benchmark. Fairness metrics inherit that noise.
1 in 4
Images that change fairness stratum entirely depending on which label source is trusted (23.2% band-crossing, n = 15,230).
17–18%
How much label noise shrinks measured disparities — real gaps look smaller than they are. We correct for it.
Six faces across the full range of skin tones, Fitzpatrick I through VI, with AI analysis mapping overlaid on each
Every audit evaluates performance across the full range of skin tones — Fitzpatrick I through VI, or Monk 1–10.

The audit most vendors have never had — because it didn't exist

Conventional bias audits compute fairness metrics on top of skin-tone labels assumed to be correct. Our peer-review-track research showed those labels disagree more often than they agree — which means conventional audits can pass a biased model, or fail a fair one. We measure the reliability of the labels first, then the performance of your model on defensible strata, with confidence intervals on everything.

Published & preregistered

Preprint: DOI 10.21203/rs.3.rs-10670899/v1
Preregistration: OSF.IO/FMSNE
Archived code: Zenodo 22039767

No model access required

We never see your code, weights, or images. Tier 1 needs only your dataset's labels; Tier 2 needs only prediction outputs. Your IP never leaves your building.

Version-locked & reproducible

Every number is computed by a release-tagged engine with SHA-256 provenance on the code, the inputs, and the report. Your audit can be re-run, byte for byte.

Nurse-founded, SBA WOSB-certified

Founded by Niya D. Pennie, MSN, RN. Certified by the U.S. Small Business Administration as a Woman-Owned Small Business — eligible for federal set-aside and supplier-diversity procurement. Audit discipline aligned to FDA change-control guiding principles.

SBA WOSB Certified
Radiant Revive audit engine: ROC curves, calibration plots, per-subgroup performance and reliability across skin-tone groups, with version-controlled commits
Fairness, calibration, reproducibility, performance — evaluated per skin-tone stratum by a version-locked engine, with SHA-256 provenance on every run.

Three engagement tiers

Fixed price, fixed timeline, written scope before you sign. Start where your data is.

Tier 1

Dataset Reliability & Representation Audit

$9,500
2–3 weeks
  • Skin-tone label reliability: agreement, weighted κ, band-crossing rate
  • Representation and positive-case counts by stratum
  • Precision ceilings: the tightest claim your data can ever support
  • Written report + findings call
You provide: dataset labels/metadata only. No images, no PHI, no model.
Tier 2 · Most Popular

Stratified Performance & Calibration Audit

$35,000
4–6 weeks
  • Everything in Tier 1
  • Per-stratum sensitivity, specificity, AUROC with confidence intervals
  • Calibration error and slope by skin tone; bootstrap disparity CIs
  • Label-sensitivity analysis under dual stratifications
  • Remediation guidance + readout call
You provide: Tier 1 inputs + model prediction outputs (probabilities). Still no code, weights, or images.
Tier 3

Preregistered Independent Validation

From $85,000
8–10 weeks
  • Everything in Tier 2, under a publicly preregistered protocol
  • Locked evaluation set held by us — opened once
  • The strongest independent-evidence artifact short of a regulatory submission
  • Citable report for buyers, partners, and diligence
You provide: Tier 2 inputs + agreement on the prespecified protocol.

Add-ons — for teams that have to file the evidence, not just read it

Add-on

Regulatory Evidence Package

+$7,500
On Tier 2 or Tier 3
  • Crosswalk of every audit finding to FDA GMLP principles, PCCP elements, ONC HTI-1 DSI source attributes, and EU AI Act Art. 9/10/13/15 obligations
  • Labeling-ready subgroup performance statements
  • Provenance appendix formatted for submission
Documentation product on the same version-locked report. Not a compliance determination.
Add-on

Continuous Monitoring

30% of tier price / re-run
Quarterly, or on each model update
  • Re-execution on the locked evaluation set, byte-for-byte reproducible
  • Version-to-version deltas per stratum with bootstrap CIs
  • Change log that slots directly into a PCCP monitoring record
Models change; your equity evidence shouldn't go stale.

How it works

Engineered so your legal and engineering teams say yes quickly.

Scoping call

30 minutes. You describe the product and data; we confirm the tier and send a fixed written quote within 24 hours.

Papers & deposit

Mutual NDA + engagement letter. No data moves before both signatures.

Secure intake

You export one CSV to our intake spec via encrypted upload. We return a SHA-256 manifest of exactly what we received.

The audit

Version-locked engine, fixed seeds, full provenance. Validation issues are flagged back to you before analysis — never silently patched.

Readout & report

Findings call first, then the report: methods, tables, confidence intervals, limitations. Your data destroyed 90 days after delivery, confirmed in writing.

If your model performs equitably, be able to prove it.

Health systems, buyers, and regulators are starting to ask for skin-tone equity evidence. The vendors with an independent audit in hand will win those conversations.

Book a Scoping Call

Radiant Revive, LLC · Texas woman-owned small business · (469) 213-8799 · info@radiantrevivemedspa.com
Independent research-grade auditing. Reports state methods, limitations, and confidence intervals; not a certification, clinical-performance claim, or regulatory determination.