As a Research Scientist, you will be responsible for improving the factuality of the Gemini model family. You will bridge the gap between research and production. You will also design advanced training methodologies to reduce hallucinations and build automated evaluation systems to measure the model's accuracy across real-world knowledge domains.

Develop evaluation methods for detecting factual inaccuracies in all LLM generated content.

Design and implement advanced pre-training, supervised fine tuning (SFT), and reinforcement learning strategies to maximize the generation of factual content.

Innovate on algorithmic interventions and data curation strategies to improve model calibration, ensuring Gemini accurately identifies its own knowledge boundaries.