What Ayush demonstrated

Built a leakage-aware turbine yield predictor end to end: framed TEY as a regression target, used only upstream sensor inputs, trained a LinearRegression baseline and a RandomForest model, and evaluated on a locked-away later time period. The candidate reported held-out MAE, RMSE, and R2, inspected residual behavior, and documented practical limits and monitoring needs in a model card.

• Leakage-Free Feature Selection
• Held-Out Model Evaluation
• Residual-Based Error Analysis