Interpretable-by-Design GAM for an Insurer's Claims Triage
Overview
What this challenge is about.
You receive an anonymized claims dataset (around 60,000 claims, target: log reserve), a feature schema (22 features), and an existing LightGBM baseline (held-out R^2 of 0.78). Train an Explainable Boosting Machine (EBM, a modern interpretable GAM) on the same features, tune via cross-validation, and produce per-feature shape plots. Compare EBM vs. LightGBM on held-out R^2, calibration by reserve decile, and actuarial sniff-test (you produce a per-feature shape-plot deck). Write a 2-page memo on the accuracy/interpretability trade-off and recommend a path forward.
The Brief
What you'll do, and what you'll demonstrate.
Quantify the accuracy/interpretability trade-off between an Explainable Boosting Machine and a black-box LightGBM for claims-reserve estimation.
Earning criteria — what you'll demonstrate
- Train and tune Explainable Boosting Machines as interpretable-by-design models
- Compare interpretable vs. black-box models fairly on regression
- Evaluate calibration alongside accuracy
- Communicate model trade-offs to an actuarial audience
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.
Career paths this builds toward
Canonical rolesData Scientist
Owning an interpretability-vs-accuracy comparison and writing the actuarial memo is exactly the day-one job of a data scientist in regulated insurance modeling.
This challenge sharpens
- generalized-additive-models
- calibration
- model-comparison
AI Safety Researcher
Pushing for interpretable-by-design models over black-box alternatives is the kind of design choice AI safety researchers advocate for in regulated industries.
This challenge sharpens
- ebm
- interpretability
- calibration
Machine Learning Engineer
Shipping calibrated regression models with reproducible training scripts and per-feature explanations transfers directly to MLE roles on regulated-modeling teams.
This challenge sharpens
- interpretability
- python
- model-comparison