Carl Ble

Research scientist at Google DeepMind interested in efficient and reliable language models, reasoning, and model security. His personal site also highlights prior work on planning, geometry, and representation learning.