Weijie Xu

I am a Senior Applied Scientist at Amazon working at LLM Evaluation and LLM Agent. Currently, I own a production system focused on domain-specific LLM evaluations within Amazon.

My experience at Amazon includes developing and implementing a unified tabular deep learning framework, successfully addressing internal prediction tasks such as safety assessments and headcount forecasting. Additionally, I have built a weakly-supervised topic modeling system capable of analyzing millions of internal documents, as well as an advanced agent-based system designed to replace traditional surveys for internal applications. My work has resulted in publications in prestigious academic conferences such as ACL and EMNLP.

Prior to Amazon, I served as a Senior Data Scientist at Schlumberger, where I co-founded the company's first data science team dedicated to addressing employee attrition challenges through innovative tabular deep learning techniques.

I hold a Master of Arts in Statistics from the University of California, Berkeley (2018), and a Bachelor of Science in Applied Mathematics with a minor in Computer Science from the University of Toronto (2017).