Candidate Matching AI Agent Across 100 Industries

This role is designed to compare documented candidate evidence with role requirements. The target deliverable is a ranked review queue with evidence. A useful evaluation should track match explanation quality, recruiter corrections and review throughput, plus human correction rate, exception rate, latency and total cost per accepted output.

Industry context changes the sources, permissions and review boundary, but the underlying task definition should remain explicit enough to compare implementations fairly.