Research Engineer, Alignment Science

Anthropic is seeking a Research Engineer for their Alignment Science team to contribute to exploratory experimental research on AI safety. The role involves building and running machine learning experiments to understand and steer powerful AI systems, with a focus on risks from future systems. Key responsibilities include testing safety techniques, running multi-agent reinforcement learning experiments, building evaluation tools, and contributing to research papers and talks. The ideal candidate has significant software, ML, or research engineering experience, familiarity with technical AI safety research, and a collaborative work style. Experience with LLMs, reinforcement learning, and complex codebases is a plus. The position offers competitive compensation, including salary (£230,000 — £515,000 GBP), equity, and comprehensive benefits. Anthropic values diversity and encourages applications from underrepresented groups. The company operates on a hybrid work model with at least 25% in-office time, preferably in the Bay Area, and offers visa sponsorship. Anthropic is committed to big science AI research, working as a cohesive team on large-scale efforts to advance steerable, trustworthy AI.