Overview
What this challenge is about.
Fit a 5-state HMM to diabetic patient data, assign latent paths with Viterbi, and validate risk monotonicity. You get a verifiable certificate.
The Brief
What you'll do, and what you'll demonstrate.
Fit a Hidden Markov Model on diabetic claim histories whose latent states correlate meaningfully with future hospitalization risk.
Earning criteria — what you'll demonstrate
- Apply the Baum-Welch (EM) algorithm to fit an HMM on real categorical data
- Use the Viterbi algorithm to decode latent state sequences
- Validate an unsupervised temporal model against a held-out clinical label
- Translate latent-state output into clinically meaningful language
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
Fitting a latent-state temporal model on real claims data and translating it for clinicians is the textbook day-one task for a junior healthtech data scientist.
This challenge sharpens
- hidden-markov-models
- em-algorithm
- model-validation
Applied AI Scientist
Choosing and validating a probabilistic temporal model on a real outcome label, then communicating to a non-technical care team, is the rhythm of applied AI work in healthtech.
This challenge sharpens
- hidden-markov-models
- time-series-modeling
- clinical-communication
Machine Learning Engineer
Productionizing unsupervised models with held-out validation and reproducible configs is the MLE craft this challenge rehearses end to end.
This challenge sharpens
- python
- model-validation
- time-series-modeling