Vietnamese-Speaking LLM Prompt Evaluation Specialist

You will create and assess prompts for large language model systems. Your work will help test how well AI handles information retrieval, reasoning, calculations, and instruction following.

You will use business documents, structured data, and related source materials to create realistic evaluation tasks. You will also review generated responses, find errors or inconsistencies, and apply project guidelines consistently.

• Write clear, accurate prompts for LLM-based systems.
• Design tasks that test retrieval, reasoning, calculations, and instruction following.
• Rate AI-generated responses against defined quality criteria.
• Identify errors, inconsistencies, and ways to improve model performance.
• Build evaluation examples from business documents, structured data, and other source materials.

What it pays and takes

The role is listed as entry level, but it requires substantial education and professional experience. Pay is not specified in the listing.

• Pay: Not specified.
• Schedule: 20 or more hours per week, with flexible working hours.
• Work arrangement: Remote and open worldwide.
• Engagement: Part-time contractor role.
• Language: Professional-level Vietnamese and strong English communication and comprehension.
• Education: Master's degree or higher.
• Experience: At least three years of professional experience.
• Core skills: Analytical and critical thinking, close attention to detail, and the ability to solve complex, multi-step problems.
• Preferred experience: Large language models, prompt engineering, AI training or evaluation, human feedback, data annotation, generative AI, content quality, or model evaluation.
• Helpful background: Film studies, cinematography, media production, or content across different genres.

How it works

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

About AI training work

AI training work is the human work behind systems that generate and understand text. People write prompts, review model responses, and label examples so AI systems can follow instructions and produce more accurate results.