Research Scientist, SAMBA, DeepMind

As a Research Scientist on the Sociotechnical Analysis of Model Behavior (SAMBA) team in the Responsible Research Group (R2), you will conduct research on model behaviors that impact people and society. You will design tools and methods for measuring the impact of AI on society, broadly defined, design experiments, and develop measures to characterize model properties relevant to responsibility and safety. As one part of your work, you will collaborate closely with other Research Scientists and Research Engineers to build frontier evaluations, apply them in practice, and publish results.

You will be joining a specialized Responsibility team within Google. While our reporting structure sits within Google's broader Trust & Safety organization, our work remains deeply integrated with DeepMind.

Artificial intelligence will be one of humanity’s most transformative inventions. At Google DeepMind, we are a pioneering AI lab with exceptional interdisciplinary teams focused on advancing AI development to solve complex global challenges and accelerate high-quality product innovation for billions of users. We use our technologies for widespread public benefit and scientific discovery, ensuring safety and ethics are always our highest priority.

Responsibilities

• Identify new research areas and questions.
• Design and conduct research and experiments.
• Develop and test new metrics and measures.
• Perform data analysis to synthesize and visualize research findings and evaluation results.
• Lead writing results for research publications.

Minimum qualifications:

• PhD in Computer Science, Economics, Computational Social Science, Data Science, or Political Science or a related field, or equivalent practical experience.
• 2 years of experience with statistical concepts, data analysis, and visualization.
• 2 years of experience with programming languages (e.g., R, Python), relevant libraries (e.g., Pandas, NumPy, Scikit-learn), or other scripting languages.
• 2 years of experience with AI evaluation.

Preferred qualifications:

• Experience creating AI evaluations or benchmarks.
• Advanced understanding of statistical concepts, and ability to perform data analysis and visualization.
• Excellent communication and collaboration skills and ability to effectively communicate and work with researchers from a wide variety of backgrounds in research publications, policy briefs, or presentations targeted at both technical and non-technical audiences.
• Peer reviewed publication on AI auditing/evaluation, especially if it relates to societal impact.