Italian Trust & Safety Data Trainer (Remote, $40/hr)

About the role

We are looking for a native‑level Italian speaker to join our Trust & Safety team on a flexible, project‑based contract. You will evaluate AI‑generated content, ensure compliance with safety policies, and help improve model behavior across Italian and English language streams.

Key responsibilities

• Review and assess AI‑generated safety‑related content for accuracy, policy compliance, and reasoning quality.
• Label and quality‑check Trust & Safety datasets covering hate speech, harassment, self‑harm, violence, misinformation, malicious activity, and other harmful content.
• Conduct red‑teaming and adversarial testing to uncover unsafe outputs, policy loopholes, and edge‑case vulnerabilities.
• Compare and rank AI responses based on safety alignment, clarity, severity, and adherence to moderation guidelines.
• Document safety risks, escalation patterns, policy gaps, and mitigation recommendations using structured rationales.
• Apply and localize Trust & Safety policies for Italian and English content while preserving linguistic and cultural accuracy.

Required profile

• Bachelor’s degree or higher in Communications, Linguistics, Psychology, Law/Policy, Security Studies, or related field (or equivalent professional experience).
• Native or near‑native proficiency in Italian and strong English communication skills (C1 or above).
• Professional experience in Trust & Safety, content moderation, risk operations, compliance, investigations, or policy enforcement.
• Proven experience with LLM red‑teaming or adversarial testing and documentation of mitigation strategies.
• Ability to work with explicit, disturbing, or psychologically sensitive content while maintaining professionalism.

Required skills

• Familiarity with AI tools and platforms such as ChatGPT, Gemini, Perplexity.
• LLM red‑teaming.
• Adversarial testing.

What we offer

• Flexible, project‑based contract work.
• Competitive hourly rate up to $40/hr.
• Remote work environment.