Gemini
Gemini
2023As a Staff Research Engineer at Google DeepMind, I am part of a research engineering team supporting the development of Gemini, Google's family of multimodal AI models. I was involved end-to-end in the post-training of the Gemini 1.0 Pro and Ultra models focusing on SFT, RLHF, and eval/deployment. Since then, I have been building post-training infrastructure with a focus on distributed reinforcement learning, driving performance and modelling improvements for scaling, coding, and agentic post-training efforts. Most recently I was directly involved in the training of the Gemini 4.0 series.