Overview
What this challenge is about.
Fit a GP regression to wind farm SCADA data, flag underperforming turbines using credible bounds, and validate results. Earn your verifiable certificate.
The scenario
The asset manager (about 80 staff, EUR 4.2B assets under management) currently runs fleet-wide alerts on raw deviations from the manufacturer curve; the false-positive rate has trained operators to ignore them, costing roughly 1.5 percent of annual energy production.
The Brief
What you'll do, and what you'll demonstrate.
Use per-turbine Gaussian Process regression to flag genuine underperformance with calibrated uncertainty and a defensible false-positive budget.
Earning criteria — what you'll demonstrate
- Apply Gaussian Process regression with composed kernels to real time-series data
- Tune kernel hyperparameters via marginal likelihood maximization
- Use credible bounds to define a defensible anomaly threshold
- Communicate GP-based decisions to a non-statistician asset team
Program Fit
Where this fits in your program.
Sharpens the same skills your degree expects you to demonstrate.
Aligned coursework coming soon.
Skills
Skills you'll demonstrate.
Each one shows up on your verified credential.
- Gaussian Processes
Apply gaussian processes to solve real industry problems and demonstrate production-level capability.
- Kernel Methods
Apply kernel methods to solve real industry problems and demonstrate production-level capability.
- Uncertainty Quantification
Apply uncertainty quantification to solve real industry problems and demonstrate production-level capability.
- Anomaly Detection
Apply anomaly detection to solve real industry problems and demonstrate production-level capability.
- Python
Write clean, efficient Python for data processing, automation, and backend services.
- Asset Management
Apply asset management to solve real industry problems and demonstrate production-level capability.
Careers
Career paths this challenge builds toward
Completing this challenge demonstrates skills that transfer directly to these roles:
Data Scientist
Probabilistic regression with calibrated bounds on industrial sensor data is the daily work of data scientists at energy and infrastructure firms.
This challenge sharpens
- gaussian-processes
- uncertainty-quantification
- anomaly-detection
Applied AI Scientist
Choosing kernels, validating coverage, and translating credible bounds into a flagging threshold is the rhythm of applied AI in operational settings.
This challenge sharpens
- gaussian-processes
- kernel-methods
- uncertainty-quantification
Machine Learning Engineer
Productionizing a per-asset GP pipeline with reproducible artifacts and a flagging report is core MLE work in industrial AI.
This challenge sharpens
- python
- kernel-methods
- anomaly-detection