Role OverviewMercor is seeking an AI Safety Expert - Red Team to perform jailbreaks, prompt injections, misuse cases, and bias exploitation on conversational AI models and agents. The ideal candidate will have native fluency in English and Portuguese, prior red teaming experience, and strong communication skills. The role involves working independently and asynchronously to improve AI model performance and ensure safety.
What You Will Do
The AI Safety Expert - Red Team will perform conversational AI model testing, generate high-quality human data, and document reproducibly by producing reports, datasets, and attack cases. They will also apply structure by following taxonomies, benchmarks, and playbooks to maintain consistent testing.
Why It Might Be a Fit
The ideal candidate will have experience in Adversarial ML, cybersecurity, and socio-technical risk expertise. They will also have strong communication skills to explain risks clearly to technical and non-technical stakeholders.
Requirements
- Native fluency in English and Portuguese (global, excluding Brazilian Portuguese)
- Prior red teaming experience in AI adversarial work, cybersecurity, or socio-technical probing
- Strong communication skills to explain risks clearly to technical and non-technical stakeholders
]]>