Researcher, Interpretability

OpenAI is seeking a Researcher in Interpretability to study internal representations of deep learning models. The role involves developing and publishing research on techniques for understanding model representations, engineering infrastructure for studying model internals at scale, and collaborating across teams to ensure the safety of powerful AI systems.

Skills and focus areas

• interpretability
• deep networks
• engineering
• quantitative reasoning
• research process
• mechanistic interpretability
• ai safety
• machine learning
• computer science