Belgian Dutch AI Safety Evaluator

About OpenTrain

OpenTrain is the #1 platform for finding and building careers in AI training and data labeling. It helps people discover projects, build a profile that reflects their experience, and apply to meaningful AI work in minutes.

• Create an OpenTrain account for free.
• Build a lasting portfolio of AI training and evaluation experience.
• Apply for flexible remote projects that match your skills.

About AI Safety Evaluation

AI training is the human side of building artificial intelligence. Evaluators review prompts and model interactions, apply clear guidelines, and provide structured judgments that help AI systems respond more safely and appropriately.

This work is part of a fast-growing technology field. Contributors help shape how advanced models handle sensitive topics, adversarial requests, and potentially harmful or dual-use information.

• Remote work using a computer and internet connection
• Flexible part-time work alongside other commitments
• A direct role in improving the safety of advanced AI systems

The Role

OpenTrain is recruiting a Belgian Dutch AI Safety Evaluator based in Belgium. You will use native or near-native Belgian Dutch fluency, cultural judgment, and strong written reasoning to evaluate how AI models handle sensitive topics.

This is an entry-level contractor opportunity, and prior AI or machine learning experience is not required. The default commitment is about 7 hours per week, with total work below 20 hours per week.

• Remote contractor role for candidates based in Belgium
• Part-time commitment of about 7 hours per week
• Pay of $48-$52 per hour
• Contractor engagement

What You’ll Do

You will create nuanced, expert-level prompts in Belgian Dutch across sensitive subject areas and assess prompts and conversations using structured project guidelines. Your evaluations will help identify unsafe behavior and improve model performance.

The work requires consistent decisions, clear documentation, and careful attention to how language can signal adversarial intent or escalation.

• Write expert-level prompts in Belgian Dutch
• Classify prompts and conversations according to structured guidelines
• Identify adversarial phrasing and escalation patterns
• Evaluate sensitive and dual-use information with sound judgment
• Document the reasoning behind your evaluations
• Maintain careful and consistent review quality

Requirements

You should be comfortable making nuanced judgments about written content and explaining those judgments clearly in English. A bachelor's degree, completed or in progress, is required.

• Native or near-native fluency in Dutch as spoken in Belgium
• Business-level written English
• Based in Belgium
• Strong written reasoning and attention to detail
• Ability to write nuanced prompts and classify text using guidelines
• Sound judgment around sensitive and dual-use information
• Bachelor's degree completed or in progress

Helpful Background

Experience reviewing, grading, or red-teaming written or technical content is helpful but not required. Backgrounds in trust and safety, content moderation, policy evaluation, or adversarial testing may translate well to this work.

• Written or technical content review
• Grading or evaluation
• Red-teaming
• Trust and safety
• Content moderation
• Policy evaluation
• Adversarial testing

How to Apply Through OpenTrain

Create or update your free OpenTrain profile, highlight your Belgian Dutch fluency and relevant evaluation experience, and apply through OpenTrain. Your profile can become part of a longer-term portfolio as you build experience in AI training and data labeling.

• Confirm that you are based in Belgium.
• Showcase your Dutch and English language skills.
• Describe relevant reviewing, moderation, evaluation, or red-teaming experience.
• Apply for this flexible AI safety evaluation opportunity.