The work
You will help test and improve large language models by creating difficult prompts, reviewing their responses, and explaining where they fall short. Your feedback will focus on whether responses are accurate, logical, relevant, and useful in practice.
The role uses your background in team leadership, people management, performance management, workforce planning, and operational execution to assess how well AI handles management-related topics.
- Create complex prompts that test large language model capabilities.
- Evaluate AI-generated responses for factual accuracy, reasoning, relevance, and overall quality.
- Find factual errors, logical inconsistencies, knowledge gaps, and other weaknesses.
- Design challenging test cases and edge cases that reveal model limitations.
- Write clear, detailed, evidence-based feedback to support measurable model improvement.
What it pays and takes
This is a part-time contractor role with a commitment of 20 or more hours per week. Pay details are not provided in the role information.
- Location: Open worldwide.
- Language: Professional fluency in English, including excellent written communication.
- Schedule: Part time, 20 or more hours per week.
- Experience level: Entry level.
- Education: A master's degree or higher in a relevant field.
- Experience: At least three years of professional, research, or teaching experience.
- Skills: Strong analytical and critical-thinking abilities, exceptional attention to detail, and the ability to challenge assumptions and find subtle factual or reasoning errors.
- Helpful background: Experience with large language models, generative AI, prompt engineering, AI evaluation, model testing, edge-case design, or stress-testing AI models.
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 text. Evaluators review model responses, identify problems, and provide examples that help improve accuracy, reasoning, and reliability.