Subha completed Energy and Renewable: Machine Learning for the Power Grid

Built a leakage-aware turbine yield predictor end to end: framed TEY as the target, excluded downstream CO/NOx emissions, trained a LinearRegression baseline and a RandomForest model, and evaluated them on a time-based held-out split. The final model achieved 1.532 MWh MAE, 1.937 RMSE, and 0.9833 R2 on held-out data, with a residual read and a concise model card covering intended use, limits, and monitoring.

• Leakage-Free Feature Selection
• Honest Held-Out Evaluation
• Model Comparison Against Baseline
• Plain-Language Model Communication