I am a Senior Research Scientist at Google Deepmind, co-building the Gemini family of models. I spend most of my time working on the core mechanics of LLM pre-training: data, evals, and scaling behavior, and how these three elements connect across the world’s languages. Previously, I drove research-to-production systems for Google Translate.

I completed my Ph.D. in Computer Science at the University of Maryland, College Park, advised by Marine at CLIP. During my time there, my dissertation won the Best Thesis Award from the Association for Machine Translation in the Americas. I currently serve as a Research Director at AMTA and an Industry Track Chair for NAACL 2027 in SF.

While much of my current research remains internal to Google, my public contributions are available in my google scholar or in tech reports (e.g. Gemini 2, 2.5, and 3.0).