Research Engineer / Scientist, Alignment Science

• Contribute to exploratory experimental research on AI safety with a focus on risks from powerful future AI systems, often in collaboration with Interpretability, Fine-Tuning, and Frontier Red Team. • Work across two research areas: AI Control (creating methods to ensure advanced AI systems remain safe in adversarial scenarios) and Alignment Stress-testing (creating model organisms of misalignment to understand how alignment failures might arise). • Build and run elegant ML experiments to understand and steer the behaviour of powerful AI systems. • Be based at least 25% of the time in Anthropic's London office, with occasional travel to San Francisco where the team's hub is located.

• Have significant software, ML, or research engineering experience. • Have some experience contributing to empirical AI research projects. • Have some familiarity with technical AI safety research. • Prefer fast-moving collaborative projects to extensive solo efforts. • Care deeply about the impacts of AI. • All interviews are conducted in Python — candidates should be proficient.

• Have experience authoring research papers in machine learning, NLP, or AI safety. • Have experience with LLMs. • Have experience with reinforcement learning. • Have experience with Kubernetes clusters and complex shared codebases.

• Test the robustness of safety techniques by training language models to subvert safety interventions. • Run multi-agent reinforcement learning experiments to test techniques such as AI Debate. • Build tooling to efficiently evaluate the effectiveness of novel LLM-generated jailbreaks. • Write scripts and prompts to produce evaluation questions testing models' reasoning abilities in safety-relevant contexts. • Contribute ideas, figures, and writing to research papers, blog posts, and talks. • Run experiments that feed into key AI safety efforts including Anthropic's Responsible Scaling Policy.

• Earn £260,000–£370,000 per year. • Receive visa sponsorship — Anthropic retains an immigration lawyer and makes every reasonable effort to support visa applications. • Access optional equity donation matching, generous vacation and parental leave, and flexible working hours. • Work at one of the world's leading AI safety research organisations on some of the most important problems in AI.

Anthropic is a public benefit corporation headquartered in San Francisco, with a mission to create reliable, interpretable, and steerable AI systems. The Alignment Science team conducts exploratory experimental research on AI safety with a focus on risks from powerful future systems. The London team is an extension of the broader Alignment Science effort, working on AI Control and Alignment Stress-testing in close collaboration with teams in San Francisco.