Short answer
What is AI red teaming?
AI red teaming is structured testing that looks for unsafe, weak, or unintended model behavior. Work can involve adversarial prompts, safety review, policy judgment, or domain-specific risk checks.
Context
Public listings may use the phrase differently, so applicants should check task scope, confidentiality rules, eligibility, and the current application page before applying.
Current technical evaluation examples
We do not currently label roles specifically as red-teaming work. The technical roles below are the closest current examples, so check the actual tasks before applying.
Mercor · software
CUDA Engineering Expert
$80 - $100 / hour
Micro1 · software
Video related professional for AI training
$65 - $80 / hour
Mercor · software
Risk-adjustment / HCC coding leader
$110 / hour