Research Engineer, Alignment Science

Anthropic is seeking a Research Engineer to join their Alignment Science team, focusing on creating safe and beneficial AI systems. The role combines scientific and engineering expertise to conduct exploratory experimental research on AI safety, particularly concerning powerful future systems.

The position involves working on critical projects including Scalable Oversight, AI Control, Alignment Stress-testing, and Automated Alignment Research. You'll be conducting machine learning experiments, testing safety techniques, running multi-agent reinforcement learning experiments, and building evaluation tools for LLM systems.

Anthropic operates as a cohesive team focused on large-scale research efforts, viewing AI research as an empirical science. The company values impact and collaboration, with frequent research discussions to ensure high-impact work. Their research builds on previous work including GPT-3, Circuit-Based Interpretability, Multimodal Neurons, and Scaling Laws.

The ideal candidate should have significant software, ML, or research engineering experience, familiarity with technical AI safety research, and a preference for collaborative projects. Strong candidates may have experience with LLMs, reinforcement learning, and complex shared codebases. The role requires at least 25% time in the Bay Area office.

As a public benefit corporation, Anthropic offers competitive compensation, benefits including equity donation matching, generous vacation and parental leave, flexible working hours, and a collaborative office space in San Francisco. They value diverse perspectives and encourage applications from candidates who might not meet every qualification, recognizing the social and ethical implications of their work.