Audit a Hiring-Screening Model for Demographic Bias
Overview
What this challenge is about.
You receive: (a) inference API access to the production model (black-box), (b) a 12,000-resume audit benchmark with self-declared gender and age-band labels (consented, GDPR-compliant), and (c) the company's stated business metric (predicting whether a recruiter advances a candidate to interview). Compute: demographic parity gap, equalized-odds gap, and selection-rate ratio (4/5 rule) across gender and age. Write a 6-page audit report with severity-graded findings, methodology limits, and a remediation roadmap. Brief the DPO in a 30-min walkthrough.
The Brief
What you'll do, and what you'll demonstrate.
Quantify the production hiring-screening model's bias across gender and age, with a report defensible to a regulator and a DPO.
Earning criteria — what you'll demonstrate
- Apply standard fairness metrics (demographic parity, equalized odds, selection-rate ratio)
- Interpret the 4/5 rule and EU AI Act conformity expectations
- Communicate audit findings to a DPO + regulator audience
- Reason about audit methodology limits (consented sample, label noise)
Program Fit
Where this fits in your program.
Sharpens the same skills your degree expects you to demonstrate.
Skills
Skills you'll demonstrate.
Each one shows up on your verified credential.
Careers
Roles this prepares you for.
Real titles. Real skill bridges. Pick the one closest to your trajectory.
AI Safety Researcher
Running a defensible third-party fairness audit and writing the regulator-facing report is the canonical applied AI-safety-research project at HR-tech, fintech, and healthtech companies.
This challenge sharpens
- fairness-metrics
- bias-auditing
- regulatory-analysis
Data Scientist
Bootstrap-CI reporting on fairness metrics is the analytical rigor data scientists are increasingly hired against under EU AI Act pressure.
This challenge sharpens
- fairness-metrics
- model-evaluation
- fairlearn