About

I’m a Senior Staff Research Scientist at Google DeepMind, where my current focus is pre-training data quality for Gemini.

My 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. Much of it builds on two decades of earlier work on retrieval and ranking, which is where a good deal of the machinery behind today’s systems originated. I am a member of the SIGIR Academy and a Distinguished Member of the ACM.

35,000+ citations · h-index 61 · Two books · 14 patents · 3 Test of Time awards

Selected work

Large language models. Co-author of Emergent Abilities of Large Language Models (7,300+ citations), the UL2, ExT5, and Charformer models, the Long Range Arena benchmark, and Efficient Transformers: A Survey.

Generative retrieval. Rethinking Search: Making Domain Experts out of Dilettantes (2021) made the case for replacing the classic index-and-rank pipeline with a single model, and Transformer Memory as a Differentiable Search Index (NeurIPS 2022) showed it could actually work. Together they helped establish generative retrieval as a research area; I went on to co-organize the first two Gen-IR workshops at SIGIR 2023 and 2024.

Personal search. Research on learning to rank from the sparse, biased feedback that personal search produces, recognized with the SIGIR 2026 Test of Time award. The work shipped across Gmail and Drive through more than 100 quality launches.

Evaluation and retrieval foundations. Three separate threads have aged well: Expected Reciprocal Rank, developed at Yahoo! Research and now a standard measure for graded relevance (CIKM 2021 Test of Time); similarity measures for short segments of text, developed at Microsoft Research (ECIR 2017 Test of Time); and the Markov Random Field model for term dependencies (SIGIR 2005 Best Student Paper, SIGIR 2017 Test of Time honorable mention), alongside work on the Indri search engine.

Teaching the field. Co-author of Search Engines: Information Retrieval in Practice, a widely used undergraduate and graduate information retrieval textbook that is freely available online, and author of A Feature-Centric View of Information Retrieval (Springer, 2011).

For a full list of publications, see my Google Scholar profile. For more, see my LinkedIn profile, CV (short), or CV (full).