The Work
You will help evaluate how AI systems personalize responses in Korean. The work combines prompt writing, response review, side-by-side comparison, and clear written reasoning.
You will assess whether responses use personal context correctly, sound natural, remain helpful, and avoid unsupported conclusions. You will also check relevant debug information and keep your judgments consistent across repeated tasks.
- Create and run multi-turn Korean prompts based on personal context.
- Evaluate responses for grounding, appropriate use of context, naturalness, helpfulness, and incorrect inferences.
- Compare model responses side by side and identify subtle differences in quality.
- Write evidence-based rationales that refer to specific conversation turns.
- Provide detailed annotations and follow project guidelines consistently.
What It Pays And Takes
This is remote, part-time contractor work open worldwide. The role requires at least 20 hours each week and strong Korean reading and writing ability.
- Pay: $15 per hour.
- Schedule: 20+ hours per week.
- Location: Open worldwide; work remotely.
- Language: Fluent Korean reading and writing.
- Experience level: Intermediate.
- Strong analytical thinking, attention to detail, and written communication skills.
- Ability to identify unsupported claims, forced connections, and differences in response quality.
- Experience with prompt design, AI quality evaluation, data annotation, content moderation, or a related analytical field is preferred.
- Ability to follow detailed guidelines and work independently.
- A relevant degree or equivalent experience is welcome.
How It Works
Apply on OpenTrain with your resume, then complete the application on the hiring site.
About AI Training Work
AI training is the human work behind systems that learn from examples, including writing prompts and reviewing model responses. Careful evaluators help shape how AI understands context and communicates with people, which is why analytical and language skills matter.