Overview
What this challenge is about.
You receive 90 days of impression logs (about 30 million recommendation events) tagged with content viewpoint labels (left-leaning, center, right-leaning, non-political) from an existing content-classification pipeline. Compute per-user viewpoint-diversity metrics (entropy, Gini, exposure ratio) over time. Identify whether new users converge to narrow exposure faster than long-tenured users. Recommend three concrete recommender changes (cold-start mixing, exploration boost on viewpoint, diversity reranking) with expected impact sketches. Deliver the analysis, recommendations memo, and a board-presentation deck.
The Brief
What you'll do, and what you'll demonstrate.
Quantify recommender-driven viewpoint narrowing on a civic platform and recommend three product changes with expected impact.
Earning criteria — what you'll demonstrate
- Design diversity metrics for recommender outputs
- Run a cohort-based exposure audit on impression logs
- Translate audit findings into concrete product-design changes
- Communicate audit results to non-technical board members
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
Designing and running a recommender-bias audit with a board-ready output is the entry point for AI safety researchers on trust-and-safety teams.
This challenge sharpens
- audit-methodology
- diversity-metrics
- recommender-evaluation
Data Scientist
Cohort-based exposure analysis with stakeholder storytelling is the data-scientist craft applied to a high-stakes policy question.
This challenge sharpens
- recommender-evaluation
- data-storytelling
- social-media-analytics
AI Product Manager
Turning audit findings into three product changes with expected impact is core AI PM work on consumer platforms under regulatory pressure.
This challenge sharpens
- recommender-evaluation
- diversity-metrics
- data-storytelling