AI Trust + Safety Analyst (Brazilian Portuguese Speakers Wanted to Train AI Models Part Time!)

Role overview

In this remote, hourly-paid contractor role, you will review AI-generated content and safety decisions, evaluate reasoning quality, and provide expert feedback to improve safety, accuracy, and clarity. You will assess, compare, and score responses against defined safety and policy standards.

You must be near-native or native in Brazilian Portuguese and comfortable working across Portuguese and English safety content.

Key responsibilities

• Annotate and review safety data across sensitive domains including hate & harassment, self-harm, violence, sexual content, misinformation, and malicious activity
• Conduct red-teaming and adversarial testing to surface failure modes and policy gaps
• Localise safety policy interpretation across languages while preserving severity and intent

Candidate profile

• Near-native or native Brazilian Portuguese proficiency
• Minimum C1 English proficiency
• 5+ years’ experience in Trust & Safety or related policy/safety functions
• Demonstrated LLM red-teaming experience
• High attention to detail and emotional resilience