Saanvi Singhal completed Predictive Maintenance: Industrial ML for Fault Detection

Built a compact bearing-fault detector end to end: identified periodic fault behavior, used kurtosis as the strongest time-domain indicator, trained a depth-capped decision tree on training bearings only, and evaluated it on held-out bearings with catch-rate and false-alarm metrics. Also wrote a maintenance memo that justified a 0.42 alarm threshold using the candidate’s own test numbers and explained the outer-race diagnosis from the dominant BPFO frequency.

• Leakage-Free Model Training
• Honest Held-Out Evaluation
• Cost-Based Operational Decision-Making