Najmos Sakib Tanvir completed Energy and Renewable: Machine Learning for the Power Grid

Built a leakage-free TEY predictor for a gas turbine using 8 available SCADA sensors, excluding downstream CO/NOx emissions to avoid target leakage. Validated on a time-based holdout split, compared Linear Regression against a tuned RandomForest, and reported honest held-out MAE, RMSE, and R2 with a residual/drift read. Authored a model card that explains intended use, limits, and monitoring in plain language.

• Leakage-Free Feature Selection
• Time-Based Holdout Validation
• Honest Held-Out Evaluation
• Model Risk Communication