leakage-safe-evaluation

This paper presents a leakage-safe and scheduler-aware machine learning framework for predicting grid job runtime, using CatBoost with temporal validation to enhance scheduling efficiency in distributed computing environments.