Muhammad Ibrahim completed Energy and Renewable: Machine Learning for the Power Grid

Built a TEY prediction model for a gas turbine using leakage-free sensor inputs and a time-based train/test split. Compared a LinearRegression baseline against a RandomForestRegressor, then evaluated the final model honestly on held-out later-period data with MAE, RMSE, and R2. Also authored a model card that explains intended use, excluded downstream emissions features, limitations, and a retraining trigger based on drift.

• Leakage-Free Feature Selection
• Honest Held-Out Evaluation
• Model Card Risk Communication