Frontier AI Safety Research Engineer

Google Inc.; Greater London; On-site; GBP 90,000 - 130,000; Full time

Job summary

DeepMind is seeking a highly adaptable Research Engineer for the Responsibility team in London. You will design and scale evaluation frameworks to assess the propensity of frontier language models to engage in harmful manipulation and build infrastructure to track model performance across releases.

You will work with Research Scientists, safety policy teams, and product stakeholders to translate safety research into concrete improvements, with opportunities to influence policy and product

Qualifications

• 3 years of experience in Python programming.
• Bachelor's degree in Computer Science, Machine Learning, Mathematics, or related field.

Responsibilities

• Rapidly prototype and deliver scalable engineering solutions for the Responsibility research portfolio.
• Architect and optimize training and inference pipelines to detect harmful manipulation in frontier language models.
• Develop post-training strategies to mitigate manipulation risks including deceptive persuasion and covert influence tactics.
• Collaborate with research scientists to translate safety research into robust implementations for cross-functional stakeholders.
• Build and maintain evaluation infrastructure to track model safety performance across releases.

Skills

Python

ML frameworks

C++ (preferred)

Education

Bachelor's degree in CS/ML/Math

Tools

JAX

PyTorch

TensorFlow

Job description

DeepMind is seeking a highly adaptable Research Engineer for the Responsibility team in London. You will design and scale evaluation frameworks to assess the propensity of frontier language models to engage in harmful manipulation and build infrastructure to track model performance across releases.

You will work with Research Scientists, safety policy teams, and product stakeholders to translate safety research into concrete improvements, with opportunities to influence policy and product