Biography

I am a Research Scientist at Google Research where I work on developing methods for improving the robustness, reliability, and efficiency of Machine Learning models. Specifically, I develop reliable uncertainty estimates for practical applications. For example, to control and improve adaptive computation capabilities, or to derive small and accurate prediction sets. Currently, I am mainly working on Natural Language Processing applications, focusing on improving the precision of information-related tasks. For example, improving the robustness of Fact Verification and Language Inference systems, and leveraging them for solving related classification and generation tasks.

I have also worked on Deep Learning for Computer Vision, Medical applications, Computational Chemistry, as well as other topics in NLP including static and contextual Word Embeddings, Large Language Models, Transfer Learning, Cross-Lingual, Question Answering, program synthesis, and more. Many of my projects have been featured in global media.