Description
Anthropic is hiring a Research Engineer on Alignment Science to build and run machine learning experiments that help understand and steer powerful AI systems. The role focuses on AI safety, including scalable oversight, AI control, alignment stress-testing, automated alignment research, assessments, safeguards, and model welfare. Responsibilities include training models to test safety techniques, running multi-agent reinforcement learning experiments, building evaluation tooling, writing research scripts and prompts, and contributing to papers and talks. The role requires significant software, machine learning, or research engineering experience, some empirical AI research experience, and familiarity with technical AI safety research; a bachelor’s degree or equivalent education, training, and experience is required. The annual salary is $350,000–$500,000 USD, with a hybrid policy requiring office presence at least 25% of the time and visa sponsorship available for some candidates.
