Portfolio Predictive AI Project Hours Estimation The Problem Predict project hours from historical data. Not generative AI. Feasibility depends entirely on data quality. The Approach ML time series prediction (XGBoost, scikit-learn). Mandatory data assessment before any model work. 300 to 400 hours estimated. No amount of model sophistication compensates for thin or inconsistent historical data. Project Hours Estimation 13 / 17
