The work
You will turn real scientific workflows into realistic, multi-step tasks for AI agents. The tasks should require agents to inspect scientific data, choose suitable methods, write and run code, use computational tools, troubleshoot results, and produce scientifically valid outputs.
- Design challenging workflows across the life sciences.
- Create input files, scientific datasets, instructions, constraints, and expected deliverables.
- Develop reproducible expert solutions and objectively verifiable ground truths.
- Create automated or semi-automated grading criteria for multi-step computational work.
- Validate scientific assumptions, calculations, code, intermediate outputs, and final answers.
- Make each task self-contained and executable in a controlled, network-isolated sandbox.
- Maintain quality and throughput while responding to reviewer feedback.
What it pays and takes
This is a remote contractor assignment lasting five weeks. The work requires strong research judgment, scientific programming, and the ability to turn authentic research workflows into bounded tasks that can be graded objectively.
- Pay: $280 per approved task.
- Schedule: At least 6 hours per day and 40 hours per week.
- Time zone: Four hours of overlap with Pacific Time is required.
- Location: Remote and open worldwide.
- Language: Professional fluency in English.
- Education: A Ph.D., postdoctoral training, or equivalent research experience in the life sciences.
- Programming: Strong scientific programming experience, particularly in Python.
- Analysis: Experience with multi-step computational scientific analyses and independently validating scientific reasoning and computational outputs.
- Helpful backgrounds: Bioinformatics, computational genomics, systems biology, computational neuroscience, biostatistics, computational drug discovery, computational biochemistry, structural biology, protein engineering, or computational microbiology.
- Useful experience: Scientific libraries, command-line tools, and reproducible research pipelines.
How it works
Apply on OpenTrain with your resume, then complete the application on the hiring site.
About AI training work
AI training work uses examples, evaluations, and human feedback to help artificial intelligence systems improve. This role uses advanced life sciences and programming expertise to create tasks that test whether AI agents can carry out real scientific work accurately.