Cultural Safety & Red-Teaming Specialist – Intermediate

This is a fully remote position, open to applicants in United Kingdom.

📋 Description

• Facilitate high-stakes adversarial testing sessions by utilizing inventive "jailbreaking" methods to investigate model vulnerabilities.

• Analyze regional social, political, and religious sensitivities to assist the model in avoiding cultural "blind spots" across diverse global markets.

• Compile safety findings into detailed risk reports, collaborating with Policy Teams to refine model safety measures.

• Assess the model’s ethical decision-making in "gray-area" situations, ensuring it prioritizes human safety and upholds institutional values.

⛳️ Requirements

• Education: A minimum of a Bachelor’s Degree in Ethics, Cyber Law, Cybersecurity, or a related discipline.

• English Proficiency: At least a B1 (Intermediate) level of English is required for documenting safety findings.

• Language: Must possess native-level fluency in at least one language.

• Context: In-depth understanding of the social and cultural norms relevant to your native language/region.

• Assessment: Candidates will undergo a subject-specific qualification exam and complete ID verification as part of the hiring process.

🏝️ Benefits

• Community Connection: Join a remote team within a 1M+ diverse global AI Community of contributors and experts.

• Impactful Expertise: Leverage your native language and academic background to influence real-world AI systems and products utilized by millions daily.

• Hands-on Innovation: Acquire experience in state-of-the-art AI research, evaluation, and fine-tuning techniques.

• Global Collaboration: Work alongside a network of specialists across 20+ domains and more than 500 languages and dialects.

• Flexible Engagement: Control your schedule while contributing significantly to the advancement of technology.

• Equal Opportunity: All eligible applicants will be considered for a contractual relationship without regard to race, color, religion, sex, sexual orientation, gender identity, national origin, disability, or protected veteran status.