The work
- Create and assess difficult scientific content used to test and improve large language models.
- Write challenging, multi-step biology problems designed to reveal limits in model reasoning.
- Produce thorough, step-by-step solutions based on rigorous scientific reasoning and accurate data interpretation.
- Contribute to evaluation benchmarks covering undergraduate biology through doctoral-level research topics.
- Analyze model performance in conceptual abstraction and experimental data analysis.
- Provide detailed annotations and structured feedback for model evaluation and fine-tuning.
What it takes
- Pay: No pay amount is provided for this role.
- Time: 20+ hours per week.
- Location: Remote project work open to candidates in the United States.
- Language: Professional fluency in English and clear plain-English communication.
- Knowledge: Strong theoretical and applied biology knowledge, from entrance-exam fundamentals through graduate or doctoral-level concepts.
- Skills: Analytical and research skills, scientific-claim validation, experimental-data interpretation, multi-variable problem solving, creative problem design, strong writing, and the ability to work independently.
- Equipment: A reliable desktop or laptop and high-speed internet connection.
- Helpful background: A master's, Ph.D., or postdoctoral degree in Biology, Biotechnology, Biochemistry, Molecular Biology, or a related life sciences discipline; teaching, tutoring, or explaining complex scientific concepts with clear language, diagrams, or visual aids.
- Work type: Contract and part-time.
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 and rating model responses and evaluating scientific content. Subject-matter experts help make those examples accurate and useful.