Conceptual Reasoning Capabilities Researcher
Redwood Research · Berkeley
Research instituteonsite$350k–$850k
At a glance
Highlights
- Evaluate and improve AI reasoning on philosophy and AI safety questions.
- Create datasets to measure progress on unverifiable conceptual arguments.
- Help shape research agenda to mitigate AI risk through better reasoning.
Ideal candidate
A sharp, gritty thinker who loves dissecting fuzzy philosophical arguments and cares deeply about mitigating AI risk.
Requirements
- Unusually sharp written argumentation with ability to assess logical validity of fuzzy arguments
- Willingness to grind through evaluating LLM-written arguments and writing detailed rubrics
- Deep interest in either AI risk or acausal interactions
- hiring process
- Application includes work tests and interviews, typically followed by a 4-month work trial before a full-time offer.
Lab profile
About Redwood Research
Ensures safe AI development through rigorous technical research, analysis, and strategic advising.
Known for
- Frontier language model experiments
- AI risk assessment
- Safety measure testing
- AI safety strategic advising
Posted · Last seen