youssef badaoui completed Predictive Maintenance: Industrial ML for Fault Detection

Built a rolling-bearing fault detector end to end: engineered time-domain and envelope-spectrum features, localized the fault to the outer race via BPFO dominance, trained a compact depth-1 decision tree on training bearings only, and evaluated it on held-out bearings with catch-rate and false-alarm metrics. The candidate also wrote a cost-based maintenance memo that justified the alarm threshold using their own results and a plain-language explanation for operations.

• Leakage-Free Feature Engineering
• Honest Held-Out Evaluation
• Physics-Based Fault Localization
• Cost-Based Decision Making