AI Red-Teamer (English & German)
Job Description
The global deployment of AI necessitates rigorous safety testing that accounts for linguistic and cultural nuances. Your expertise in German and English will be pivotal in uncovering subtle biases, vulnerabilities, and potential misalignments in AI systems, ensuring they operate ethically and effectively for German-speaking communities. This role directly contributes to the trustworthiness of cutting-edge AI.
Key Responsibilities
Strategically probe AI models for safety and ethical breaches, focusing on content generation and understanding in German and English.
Craft adversarial prompts designed to elicit toxic, biased, or factually incorrect responses, particularly those sensitive to German cultural or historical contexts.
Document all identified failure modes with precise examples, linguistic analysis, and recommendations for model improvement.
Participate in iterative testing cycles, providing continuous feedback to AI development teams.
Evaluate AI's ability to handle complex German grammatical structures, idiomatic expressions, and regional variations without introducing errors or biases.
Ideal Qualifications
Native or near-native proficiency in German (including strong grasp of formal/informal registers) and English.
Previous experience in linguistic quality assurance, content moderation, or cybersecurity testing.
Deep understanding of German culture, societal norms, and potential areas of sensitivity.
A keen eye for detail and exceptional analytical abilities to detect subtle AI misbehaviors.
Familiarity with prompt engineering and the operational principles of large language models.
Academic background in German studies, linguistics, or a related field is a plus.
Project Timeline
Start Date: Within 2 weeks
Duration: Ongoing, minimum 4 months
Commitment: Flexible, 10-20 hours per week
Contribute your linguistic and analytical prowess to secure the future of AI!