Abstract: The Large Enriched Germanium Experiment for Neutrinoless Double-Beta Decay (LEGEND) will deploy up to 200 kg of High-Purity Germanium (HPGe) detectors in its first stage to search for neutrinoless double-beta decay ($0\nu\beta\beta$). In this study, we present a data-driven approach to remove anomalous events captured by HPGe detectors powered by artificial intelligence (AI). We utilize Affinity Propagation (AP) to cluster signals based on their shape and a Support Vector Machine (SVM) to classify them into different categories. We train, optimize, and test the model on data taken from a HPGe detector installed in a liquid argon test stand. We demonstrate that our model gives maximum physical event sacrifice of $0.016 ^{+0.005}_{-0.004} $ \% when performing data cleaning cuts. The AP-SVM model can be applied to classification of unlabeled time-series data from a variety of sources, and is being used to accelerate data cleaning development for LEGEND-200.