Bio

Donald Metzler is a Senior Staff Research Scientist at Google DeepMind, where his current focus is pre-training data quality for Gemini.

His research is on large language models: how they are pretrained, how their outputs can be detected and evaluated, and how they store and retrieve knowledge. He is a co-author of “Emergent Abilities of Large Language Models” and of the UL2, ExT5, and Charformer models. His earlier work helped establish generative retrieval as a research area, through “Rethinking Search: Making Domain Experts out of Dilettantes” and “Transformer Memory as a Differentiable Search Index.”

Over his career at Google he has led research groups of up to thirty engineers and scientists, including the Apps Search Quality and Intelligence team, whose work on personal search shipped across Gmail and Drive.

Prior to joining Google, he was a Research Assistant Professor at the University of Southern California (USC) and a Senior Research Scientist at Yahoo! Research. He is a Distinguished Member of the ACM, a member of the SIGIR Academy, and the recipient of Test of Time awards at SIGIR 2026, CIKM 2021, and ECIR 2017, along with multiple Best Paper awards. He has served as the Program Chair of the WSDM and ICTIR conferences and on the editorial boards of ACM Transactions on Information Systems, ACM Transactions on the Web, and the Information Retrieval Journal. He has published over 100 research papers, holds 14 issued US patents, and is the author of “A Feature-Centric View of Information Retrieval” and a co-author of “Search Engines: Information Retrieval in Practice”.