Summary
Song Xu is a Machine Learning Engineer with eight years of experience building data pipelines and ML/statistics-driven predictive systems, specializing in anomaly and fraud detection across cloud and e-commerce domains. Currently at DoorDash, Song has contributed to both fraud detection for dashers and search quality work involving NLP, DNNs, indexing, and model serving. Prior roles at Microsoft and Signifyd reflect a track record of designing online/offline modeling pipelines, production microservices, and features that materially reduced fraud losses. His academic background—PhD in Biomathematics and research at Stanford and UCLA—underscores a strong theoretical and quantitative foundation applied to practical engineering problems. Comfortable bridging Python/Java production systems and research-grade statistical modeling, Song brings a rare mix of deep theoretical training and hands-on delivery that drives robust, monitored ML services.
8 years of coding experience