Research Scientist, LLM Mechanistic Interpretability

DeepMind; Greater London; On-site; GBP 120,000 - 180,000; Full time

Job summary

DeepMind is seeking a highly motivated Research Scientist to advance mechanistic interpretability of large language models, exploring self-conceptions, theory of mind, and model cognition. You will design experiments and collaborate with cross‑functional teams to push understanding of Gemini and related models.

The role emphasizes publishing findings, communicating complex ideas to diverse audiences, and contributing to safety-aligned AI research in a leading lab.

Qualifications

• PhD or equivalent research experience in a technical field.
• 3+ years designing, executing, and analyzing empirical AI research.
• Experience using ML frameworks (PyTorch, JAX) to run experiments.
• Experience with large language model architectures and internal representations.
• Publication record in mechanistic interpretability, AI safety, or ML.

Responsibilities

• Define and lead a research agenda on mechanistic interpretability of LLMs, including self-conceptions, theory of mind, and cognition.
• Adapt and extend state-of-the-art mechanistic interpretability techniques to Gemini and other models.
• Design and execute experiments to identify internal vectors influencing model outputs.

Skills

Experimental design

Data analysis

Scientific communication

Education

PhD in Computer Science, Mathematics, Physics or related field

Tools

PyTorch

JAX

Job description

DeepMind is seeking a highly motivated Research Scientist to advance mechanistic interpretability of large language models, exploring self-conceptions, theory of mind, and model cognition. You will design experiments and collaborate with cross‑functional teams to push understanding of Gemini and related models.

The role emphasizes publishing findings, communicating complex ideas to diverse audiences, and contributing to safety-aligned AI research in a leading lab.