Frontier AI Safety Research Engineer

DeepMind Technologies Limited; Greater London; On-site; GBP 90,000 - 140,000; Full time

Job summary

Google DeepMind is seeking a research-focused Software Engineer embedded within the organization to prototype and deploy large-scale safety tests. You will help design experiments, prototype safe AI systems, and collaborate with research programs across universities and institutes globally.

The role emphasizes advancing AI safety evaluation, developing scalable pipelines, and publishing results with the wider research community, while contributing to responsible AI practices within a leading AI

Qualifications

• Master's degree or PhD in Computer Science or a related technical field.
• Experience working on AI safety or responsible AI research.

Responsibilities

• Prototype and deliver scalable engineering solutions across the Responsibility research portfolio.
• Architect and optimize training and inference pipelines to detect and evaluate harmful manipulation behaviors in frontier language models.
• Develop post-training strategies to mitigate manipulation risks including deceptive persuasion, sycophancy, and covert influence tactics.
• Collaborate with Research Scientists to translate safety research into robust implementations and present results to cross-functional stakeholders.
• Build and maintain evaluation infrastructure to systematically track model safety performance across releases.

Skills

AI safety

Responsible AI

Python

C++

LLM evaluation

Robustness testing

Evaluation pipelines

Research collaboration

Education

Masters/PhD in CS/related

ML focus in degree

Tools

Python

C++

Job description

Google DeepMind is seeking a research-focused Software Engineer embedded within the organization to prototype and deploy large-scale safety tests. You will help design experiments, prototype safe AI systems, and collaborate with research programs across universities and institutes globally.

The role emphasizes advancing AI safety evaluation, developing scalable pipelines, and publishing results with the wider research community, while contributing to responsible AI practices within a leading AI