Conduct careful web and document research to develop training and benchmarking tasks for frontier AI models. Responsibilities include retrieving and verifying claims, cross-referencing academic and financial sources, tracking original sources, using research-enabled LLM tools, rapidly learning new domains, and clearly communicating reasoning, uncertainty, and findings in written and verbal formats. This is a fully remote, one-month contract requiring at least 15 hours weekly, with an expected average of 20 hours per week.