You will help improve AI systems by reviewing Nepali responses and reference content, writing or improving expert answers, and explaining your decisions in clear English. You will follow project guidelines and keep your judgments consistent across repeated review tasks.
- Write or improve responses in Nepali and explain your reasoning clearly.
- Review source materials and AI-generated responses against project guidelines.
- Identify meaning changes, unsupported claims, incorrect honorifics, unnatural translations, and culturally awkward wording.
- Give clear feedback that helps improve AI model behavior.
- Maintain consistent quality across review tasks.
What It Pays And Takes
- Pay: $15 per hour.
- Estimated project earnings: $2,400 based on the schedule and duration shown on OpenTrain.
- Time: 20+ hours per week.
- Work arrangement: Remote contractor work across time zones.
- Location: Open to candidates in Bangladesh, Bhutan, India, and Nepal.
- Education: Bachelor's degree or higher in Mathematics.
- Language: Professional fluency in Nepali, including reading and writing.
- English: Strong professional proficiency, at least C1, to follow guidelines and provide detailed feedback.
- Experience: At least 3 years in writing, editing, localization, translation, content quality assurance, journalism, education, or related language-focused work.
- Knowledge: Strong understanding of Nepali grammar, Devanagari script, spelling, punctuation, honorifics, and style.
- You should be able to handle formal and informal registers, regional and dialect variation, and different audiences, including Nepal and Nepali-speaking diaspora communities.
- Prior AI training or annotation, translation quality assurance, editorial quality assurance, or rubric-based evaluation experience is strongly preferred.
- Prior AI training experience is not required.
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 generate and understand language. People review model responses, write examples, and explain what is correct so AI systems can improve; specialized knowledge and careful judgment are especially useful for this work.
