LLM Safety Evaluator Hebrew and English

The work

You will evaluate AI-generated responses and create safety-focused content in English and Hebrew. Your work will help identify unsafe or unintended model behavior and make responses clearer and more reliable.

Some tasks may involve explicit or sensitive material.

• Curate and label adversarial or safety-sensitive training examples.
• Review and score model outputs.
• Document safety failures and policy gaps.
• Stress-test models to find unsafe or unintended responses.
• Generate evaluation content in English and Hebrew.

What it pays and takes

This is a fully remote, hourly contractor role with part-time status. The work requires at least 20 hours each week.

• Pay: $26 to $38 per hour, with an expected rate of $32 per hour.
• Work arrangement: Fully remote contractor role.
• Time commitment: 20 or more hours per week.
• Languages: English and Hebrew required.
• Work type: Part-time contract.
• Content note: Tasks may include explicit or sensitive material.

How it works

Apply on OpenTrain. The employer reviews applications there.

About AI training work

AI training is the human work behind systems that generate text and other content. People review examples, rate model responses, and explain failures so AI systems can behave more safely and usefully.