Teams at FutureSearch

Verify and quality-check evaluation tasks for frontier AI models: confirm correct answers and evidence, spot and explain errors, apply skeptical verification, and communicate concise findings. Work remotely ~20 hours/week on a one-month contract (possible extension), meet delivery deadlines, and prioritize what needs deeper review versus quick resolution.

Build and own end-to-end backend systems for forecasting and agentic LLM research: design, deploy, monitor, debug, and harden features and pipelines (benchmarks, agents, production services). Prototype quickly, work with LLM APIs and coding agents, and collaborate with researchers to improve evaluation and forecasting tools.

Conduct research to evaluate and improve forecasting and agentic LLM web research. Formulate open problems, devise innovative solutions, build evaluations, benchmarks, and datasets, publish findings, and collaborate with frontier labs to improve AI epistemics and decision-making.