I am a senior staff research scientist at Google DeepMind. I am the co-creator of Gemini-0801, Google's first #1 on LMSYS. I'm also the eval expert behind 2.5 models such as those attaining #1 on WebDev Arena and HLE. More broadly, I am a core contributor to Gemini 1, 1.5, 2, and 2.5 working across fundamentals in RL, evals, and data. I co-lead their papers and results. Prior to Gemini, my most no