Jobs

Anthropic is seeking a Senior Research Scientist for their Reward Models team to lead research in specifying and learning human preferences at scale. This role involves developing novel architectures and training methodologies for RLHF, researching LLM-based evaluation, and mitigating reward hacking to make AI systems like Claude more useful and aligned with human values. The ideal candidate will have a strong research background in reward modeling or RLHF, experience with large-scale experiments, and a passion for building highly capable and safe AI systems. This is an opportunity to drive ambitious research agendas and ship practical improvements to production systems, working on critical problems in AI alignment with access to frontier models and significant computational resources.