Portuguese AI Safety Evaluator

About OpenTrain

OpenTrain AI helps people find and build careers in AI training and data labeling. As the hiring and contracting organization for this role, OpenTrain gives contributors a way to develop credible experience and build a lasting portfolio in a fast-growing field.

• Remote contract opportunity
• Part-time commitment of about 7 hours per week
• Pay of $40-$44 per hour
• Training provided on the workflow

About AI Safety Evaluation

AI training is the human work behind modern artificial intelligence. Contributors review model outputs, write examples, classify content, and provide feedback that helps AI systems behave more accurately and safely.

Safety evaluation focuses on how models respond to sensitive topics, dual-use information, adversarial prompts, and escalating language. Your language fluency and cultural judgment will help assess model behavior in Portuguese.

• Contribute to the development of advanced AI systems
• Use structured human judgment to evaluate model behavior
• Work remotely with flexible part-time hours

The Portuguese AI Safety Evaluator Role

OpenTrain is recruiting a Portuguese AI Safety Evaluator to help improve the safety of advanced AI models. You will use Portuguese fluency, cultural judgment, and strong written reasoning to assess sensitive conversations and document consistent decisions.

This role is suitable for a careful contributor who can work thoughtfully with sensitive and dual-use information. Prior AI or machine learning experience is not required.

• Native or near-native Portuguese fluency
• Business-level written English
• Entry-level opportunity with training provided
• Worldwide remote applicants are welcome
• Applicants in Portugal and wider Western Europe may be a strong match

What You’ll Do

You will create expert-level prompts in Portuguese across sensitive subject areas and evaluate how models respond. You will follow structured guidelines, identify concerning patterns, and explain the reasoning behind your judgments.

• Write expert-level prompts in Portuguese
• Classify prompts and conversations using structured guidelines
• Flag adversarial phrasing and escalation patterns
• Identify sensitive or dual-use information
• Document the reasoning behind evaluation decisions
• Provide written feedback to help strengthen model behavior

Requirements and Helpful Background

Applicants must be able to make sound, consistent decisions when handling sensitive content. A bachelor's degree, either completed or currently in progress, is required.

Experience reviewing, grading, or red-teaming written or technical content can be helpful. Backgrounds in trust and safety, content moderation, policy evaluation, or adversarial testing are also relevant, but previous AI or machine learning experience is not required.

• Strong written reasoning
• Close attention to detail
• Sound judgment with sensitive and dual-use information
• Ability to follow classification guidelines
• Ability to document decisions clearly
• Bachelor's degree completed or in progress

Work Details

This is a remote, part-time contractor opportunity with a default commitment of about 7 hours per week. Compensation is $40-$44 per hour, and training is provided on the workflow.

• Employment type: Contractor and part time
• Data type: Text
• Primary work: Classification and red teaming
• Languages: Portuguese and English
• Time requirement: Less than 20 hours per week
• Compensation: $40-$44 per hour

How to Apply Through OpenTrain

Create a free OpenTrain account to apply for this opportunity and build your AI training profile. OpenTrain helps contributors discover work that matches their language skills and experience while developing a portfolio for future AI training projects.

• Create your free OpenTrain account
• Highlight your Portuguese fluency, English writing ability, and relevant review experience
• Apply in minutes through OpenTrain
• Use the role to begin building experience in AI safety evaluation