About
I study natural language processing, generative AI, and cognitive science, focusing on building practical systems that make information universally accessible.
Current focus — Eval and Factuality: How do we define quality, and how can we quantify quality using models, to provide training signals and evaluation. I have done related work in factuality and grounding, in expertise, uncertainty/hedging, and aesthetics of multimodal output.
Before joining Google DeepMind, I was a tenured Associate Professor at Penn State, where I led a research group in cognitive science and NLP, and previously worked on cognitive models at Carnegie Mellon University. I hold a PhD in Informatics from the University of Edinburgh. If you are interested in post-PhD collaborations on NLP and dialogue, please apply directly to Google DeepMind.
Earlier in life, I worked for a newspaper and a radio station in Berlin, and have called Mainz, Berlin, Dublin, Edinburgh, and Pittsburgh home before settling in New York. I support, as a development trust board member, the University of Edinburgh, and the Jazz Gallery, New York City. Outside of research, I am an instrument-rated pilot—as Charles Lindbergh put it, "Science, freedom, beauty, adventure: what more could you ask of life?"
Selected Publications
- Gemini 2.5: Pushing the Frontier with Advanced Reasoning, Multimodality, Long Context, and Next Generation Agentic Capabilities
- Gemini 1.5: Unlocking Multimodal Understanding Across Millions of Tokens of Context
- Investigating Content Planning for Navigating Trade-offs in Knowledge-Grounded Dialogue
- Gemini: A Family of Highly Capable Multimodal Models
- Measuring Attribution in Natural Language Generation Models
- Evaluating Attribution in Dialogue Systems: The BEGIN Benchmark