Overview
What this challenge is about.
You receive 36 months of weekly marketing-spend and outcome data for 8 sample brands. Build a per-brand baseline gradient-boosting MMM model, then build two more base learners (a Bayesian Ridge regression with adstock + saturation transforms, and a small TabNet or MLP). Use proper time-series CV (rolling origin), then stack with a meta-learner that respects the chronological ordering. Report uplift over baseline by brand and a robustness metric (variance of predicted ROI across CV folds). Present the result in a 12-slide deck for a mixed analytics/marketing audience.
The Brief
What you'll do, and what you'll demonstrate.
Determine whether a stacked ensemble of three model families improves marketing-mix-modelling robustness over a gradient-boosting baseline.
Earning criteria — what you'll demonstrate
- Build and evaluate a stacked ensemble across heterogeneous model families
- Apply rolling-origin cross-validation correctly for time-series problems
- Quantify robustness, not just point accuracy, of marketing-mix predictions
- Translate model results into business-relevant ROI uplift
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.
Applied AI Scientist
Designing stacked ensembles that improve a real production metric is the applied-AI-scientist's signature deliverable in marketing or analytics teams.
This challenge sharpens
- ensemble-methods
- stacking
- model-evaluation
ML Researcher
Comparing heterogeneous learners with time-series CV and Bayesian components is the kind of rigour ML-research interviewers probe for.
This challenge sharpens
- bayesian-regression
- time-series-cv
- ensemble-methods
Data Scientist
Stacking pipelines and per-channel robustness reporting are skills senior data-scientist roles require for any attribution or causal-impact team.
This challenge sharpens
- regularization
- stacking
- model-evaluation