AI Personalization Evaluator (Vietnamese)

You will assess whether a conversational AI feature uses personal context from earlier conversations and connected activity in a relevant, accurate, and helpful way. Your work will combine Vietnamese language judgment, prompt design, and structured feedback for AI training.

You will create multi-turn prompts based on personal experiences, review model responses, compare two answers side by side, and write clear reasons for your ratings. You will also check debug information and delete evaluation conversations when required.

• Create and run conversational prompts that use personal context across about one to five turns.
• Check whether personalization follows the prompt and is supported by evidence.
• Identify incorrect personalization, weak inferences, hallucinations, forced connections, and robotic overnarration.
• Compare responses for helpfulness, ease of use, naturalness, and enjoyment.
• Write concise explanations that cite specific conversation turns and support your rankings.
• Review summaries and relevant data sources in debug information to confirm they were used correctly.
• Delete evaluation conversations as required for careful data handling.

This is a remote contractor engagement focused on Vietnamese-language evaluation. You must have access to a desktop or laptop, reliable internet, a primary personal Google account, and relevant personal data sources that you are willing to enable for the evaluation.

• Pay: $15 USD per hour.
• Schedule: 20+ hours per week, with full-time availability required.
• Availability: At least 4 hours per day and up to 40 hours per week.
• Time-zone overlap: 4 hours with Pacific Time.
• Length: 3 months.
• Language: Strong Vietnamese reading and writing ability.
• Experience level: Intermediate.
• Required: Strong judgment on nuanced or unclear AI responses and the ability to create creative multi-turn prompts from personal context.
• Required: Understanding of personalization quality, grounding, integration, and helpfulness.
• Required: Clear, structured writing; precise references to conversation turns; independent work habits; and strong communication.
• Preferred: A degree or equivalent experience in policy, law, ethics, linguistics, journalism, computer science, or a related analytical field.
• Preferred: Experience in data annotation, AI quality evaluation, content moderation, or similar review work.

Apply on OpenTrain with your resume, then complete the application on the hiring site.

AI training is the human work behind systems that learn from examples, including reviewing model responses and rating which answers are more useful. People with strong language skills and careful judgment help make these systems more accurate, natural, and reliable.