Korean Trust & Safety Data Trainer | $50/hr Remote

Korean Trust & Safety Data Trainer Work Snapshot

• Job Type: Contract
• Location: Remote
• Compensation: $20-$50 per hour
• Level: Middle to Senior Level

Roles & Responsibilities

• Review AI-generated content and moderation decisions for safety, correctness, reasoning quality, and policy compliance
• Label and quality-check Trust & Safety datasets involving hate speech, harassment, self-harm, violence, misinformation, malicious activities, and harmful content
• Perform adversarial testing and red-teaming to identify policy loopholes, unsafe outputs, exploit patterns, and edge-case vulnerabilities in AI systems
• Evaluate and compare multiple AI responses based on safety alignment, contextual appropriateness, and clarity of reasoning
• Analyze Korean and English content with attention to cultural nuance, slang, coded language, implicit threats, and contextual meaning shifts
• Apply and localize Trust & Safety policies consistently across multilingual content while documenting escalation paths and moderation rationales
• Produce concise analytical explanations explaining why outputs are safe, unsafe, ambiguous, or policy-violating
• Support AI model improvement through annotation workflows, safety evaluations, quality assurance reviews, and structured documentation

Requirements

• Education: Bachelor s degree or higher in Communications, Linguistics, Psychology, Law/Policy, Security Studies, or a related field; equivalent professional experience may also qualify
• Native or near-native Korean proficiency with strong English skills (C1 or higher) for bilingual moderation and policy interpretation
• Professional experience in Trust & Safety, content moderation, policy operations, investigations, compliance, risk assessment, or related safety-focused work
• Proven LLM red-teaming or adversarial testing experience with the ability to identify edge cases and recommend mitigation strategies
• Strong understanding of safety domains including hate & harassment, self-harm, violence, misinformation, malicious activity, bias, and unsafe content patterns
• Localization or translation experience preferred, especially preserving severity, tone, and intent between Korean and English contexts
• Emotional resilience and professionalism when handling explicit, violent, psychologically disturbing, or toxic content
• Excellent analytical writing skills, attention to detail, and ability to apply evolving moderation guidelines consistently
• Experience with AI systems and tools such as ChatGPT, Gemini, Perplexity, or similar platforms strongly preferred