Senior Researcher, Interpretability and AI Safety

Corehr; Oxford; Hybrid; GBP 50,000 - 70,000; Full time

Job summary

The University of Oxford’s Department of Engineering Science seeks a full-time Senior Researcher to advance interpretability and AI safety for continually learning systems in the central Oxford area. The role focuses on evaluating dynamic models, designing experiments, and documenting results to ensure safety properties hold under changing data.

You will join the Technical Safety and Governance lab and work with the Oxford Martin AI Governance Initiative, contributing to collaboration,

Qualifications

• PhD in ML, CS, or closely related field.
• Several years of research experience.
• Experience in interpretability, evaluation, lifelong learning, or AI safety.
• Strong publication record in top venues (NeurIPS/ICML/ICLR/ACL/EMNLP).

Responsibilities

• Research on interpretability and AI safety for continually learning systems.
• Design, run, and analyze large-scale experiments on foundation models.
• Collaborate with the TSG lab and Oxford Martin AI Governance Initiative.
• Publish findings and contribute to governance-focused research.

Skills

Interpretable ML

Model evaluation

Continual learning

AI safety

Publication record

Education

PhD in ML/CS or related

Job description

Senior Researcher in Interpretability and AI Safety

central Oxford

We are seeking a full time Senior Researcher to joinTechnical Safety and Governance (TSG) labat the Department of Engineering Science in central Oxford. The post is funded by the Oxford Martin AI Governance Initiative and is fixed-term to 1 year, with the possibility of an extension for an additional year.

The Senior Researcher will work on Interpretability, evaluations and AI safety for continuously learning systems. They will evaluate systems whose behaviour keeps moving, and track, from the inside, whether the structures that carry their capabilities and safety properties hold up under the change.

You will be responsible for the follow : (full details of duties available from the Job Description)

Research

Collaboration and engagement

You will have completed a PhD in machine learning, computer science, or a closely related field and with several years of experience and have research experience in at least one of: mechanistic or representational interpretability; model evaluation and benchmarking; continual or lifelong learning; or AI safety (alignment, scalable oversight, adversarial robustness). A strong publication record at relevant venues (NeurIPS, ICML, ICLR, ACL, EMNLP, or established AI safety venues and workshops) together with the ability to design, run, and make sense of large-scale experiments on foundation models.

Informal enquiries may be addressed to Fazl Barez at [e-mail].

For more information about working at the Department, see

www.eng.ox.ac.uk/about/work-with-us/

The Department holds an Athena Swan Bronze award, highlighting its commitment to promoting women in Science, Engineering and Technology.

AI, interpretability, continuous learning, evaluations chain-of-thought