Find your next role
Discover roles across ambitious teams
2,872 jobs
- More jobs
Confidential
AI Red Teaming Specialist
Remote · Nigeria · Contract · Mid level
$10 / hour
About the role
Design and test AI model tasks to identify failures and improve performance. Investigate facts and write benchmark prompts with accurate answers.
What you’ll do
• Investigate and verify facts through credible sources. • Design research tasks to test AI models. • Test AI outputs and identify failures. • Analyze model search trajectories and reasoning failures. • Write clear benchmark prompts and answer keys.
What you’ll bring
• Experience in investigative research or AI benchmark authoring. • Familiarity with AI benchmarks like GAIA and BrowseComp. • Strong web research and OSINT skills. • Ability to analyze model search trajectories and failures. • Strong writing skills for benchmark tasks.