AI Red-Team Specialist — English & Malay (Remote)

About the role

We're building a global red-teaming bench to help make AI systems safer before they reach the public. This bounty is focused on bilingual English/Malay experts who can stress-test AI model outputs, surface weaknesses, and generate the structured feedback data that trains safer, more robust models.

You'll work directly with sensitive, nuanced content — probing for bias, misinformation, and harmful or unsafe behaviors — and documenting your findings with precision. This is fully remote, flexible, text-based work suited to detail-oriented bilingual professionals who care about the responsible development of AI.

What you'll do

• Craft adversarial prompts designed to surface model vulnerabilities, biases, and unsafe behaviors
• Evaluate AI-generated outputs in both English and Malay for accuracy, safety, and appropriateness
• Identify and document instances of misinformation, harmful content, or policy violations
• Provide structured written feedback and annotations that will be used to retrain and improve model safety
• Collaborate with a distributed team of reviewers and safety researchers to refine red-teaming methodology
• Maintain strict confidentiality and consistency when handling sensitive or adversarial content

Requirements

• Native or near-native fluency in both English and Malay (written and spoken)
• Strong critical thinking skills and comfort engaging with sensitive, controversial, or adversarial content
• Excellent written communication and attention to detail
• Ability to work independently in a remote, asynchronous environment
• Prior experience with content moderation, linguistics, translation, QA, or AI training data is a plus
• Familiarity with AI/LLM concepts or prior red-teaming/data annotation experience is a strong plus

Compensation

Hourly rate: $17–$25 USD, based on experience and demonstrated performance during onboarding tasks. Flexible, remote, part-time or full-time hours available.