Research Engineer Scientist (AI)

Location: London, UK; at least 25% presence in the London office required, with occasional travel to San Francisco

Annual salary: £260,000–£370,000 GBP

Company

Anthropic is a public benefit corporation developing reliable, interpretable, and steerable AI systems intended to be safe and beneficial.

What you will do

• Design and run empirical machine learning experiments for AI alignment and safety research.
• Develop AI Control methods to keep advanced AI systems safe and harmless in unfamiliar or adversarial scenarios.
• Conduct alignment stress-testing by training language models to subvert safety techniques and studying potential alignment failures.
• Run multi-agent reinforcement learning experiments, including research on AI Debate.
• Build tooling for evaluating LLM-generated jailbreaks and create prompts and scripts for safety-relevant reasoning evaluations.
• Contribute ideas, figures, and writing to research papers, blog posts, talks, and the Responsible Scaling Policy.

Requirements

• Significant software, machine learning, or research engineering experience.
• Experience contributing to empirical AI research projects and familiarity with technical AI safety research.
• Ability to work collaboratively on fast-moving research projects and communicate effectively.
• Bachelor’s degree or an equivalent combination of education, training, and experience in a relevant field.
• Availability to work from the London office at least 25% of the time and travel to San Francisco occasionally.
• Python is used for all interviews.

Nice to have

• Research papers in machine learning, NLP, or AI safety.
• Experience with large language models and reinforcement learning.
• Experience with Kubernetes clusters and complex shared codebases.

Culture & Benefits

• Collaborative research environment focused on a small number of large-scale AI research efforts.
• Frequent research discussions and emphasis on communication and measurable impact.
• Flexible working hours and a hybrid office policy.
• Competitive compensation, optional equity donation matching, generous vacation, and parental leave.
• Visa sponsorship may be available, with immigration lawyer support.

Hiring process

• All interviews for this role are conducted in Python.
• Visa sponsorship eligibility is assessed based on the specific role and candidate.