Computer Science (STEM) Task Designer

The work

You will turn real computational life sciences practices into self-contained tasks for evaluating AI agents. Each task should require the agent to inspect data, choose methods, run code, troubleshoot results, and produce scientifically valid outputs.

You will create the files, datasets, instructions, limits, expected deliverables, and computational steps needed for each workflow. You will also build reproducible expert solutions, define objective ground truths, validate scientific and computational results, and respond to reviewer feedback.

The partner title identifies this as Computer Science (STEM) work. Review the listed duties and qualifications before applying because the assignment focuses on computational life sciences task design.

• Design multi-step scientific workflows across life sciences disciplines.
• Create realistic datasets, input files, instructions, constraints, and deliverables.
• Develop automated or partly automated grading criteria that run in a network-isolated sandbox.
• Check scientific assumptions, computational results, reproducibility, and task quality.

What it pays and takes

This is a remote contractor assignment lasting five weeks. The work requires at least 20 hours per week, including a four-hour overlap with Pacific Standard Time.

The assignment requires advanced research experience and strong scientific programming ability, especially in Python. Helpful backgrounds include bioinformatics, computational genomics, systems biology, computational neuroscience, biostatistics, computational drug discovery, computational biochemistry, structural biology, protein engineering, and computational microbiology.

• $280 per approved task.
• Remote assignment open worldwide.
• At least 20 hours per week.
• Four-hour work overlap with Pacific Standard Time.
• Ph.D., postdoctoral training, or equivalent research experience in the life sciences.
• Strong Python and scientific programming experience.
• Experience with multi-step computational scientific analysis.
• Ability to independently validate scientific and computational results.
• Ability to turn authentic research workflows into bounded, self-contained, objectively gradable tasks.
• Experience with scientific libraries, command-line tools, and reproducible research pipelines is helpful.
• This contractor assignment does not include medical or paid leave.

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, reviewing, coding, and rating model outputs. People with research and technical expertise are needed to create reliable examples and judge whether an AI system produces valid results.