Tag: machine-learning

Anomaly Detection at Scale: Graph-Aware XGBoost for 1.2M Events/sec

Building Trufflow's high-throughput anomaly detection pipeline using graph-aware feature engineering, Polars vectorization, and XGBoost with a 73x accuracy lift over baseline methods.