Computational Life Sciences AI Task Designer

You will design self-contained scientific workflows that test whether AI agents can carry out realistic computational research. The work focuses on scientific reasoning, code execution, troubleshooting, and valid research deliverables rather than simple question answering.

Tasks run in controlled, reproducible, network-isolated environments. You will create the data, instructions, constraints, expected results, and grading methods needed to evaluate each workflow objectively.

• Design challenging computational workflows across the life sciences.
• Create realistic input files, scientific datasets, instructions, constraints, and expected deliverables.
• Develop reproducible expert solutions and objectively verifiable ground truths.
• Build automated or semi-automated grading criteria for multi-step analyses.
• Validate scientific assumptions, calculations, code, intermediate outputs, and final answers.
• Package dependencies, data, and computational resources so tasks run reliably.
• Maintain quality and throughput while using reviewer feedback.

This is a remote, part-time contract role for an individual contributor. The listing does not state a pay rate.

• Pay:
  Not specified in the listing.
• Time: 20+ hours per week.
• Location:
  
  Worldwide and remote.
• Language:
  English.
• Work type:
  Contractor and part-time.
• Experience level:
  Listed as entry level, with a requirement for a Ph.D., postdoctoral work, or equivalent life sciences research experience.
• Required skills:
  
  Strong scientific programming experience, especially in Python, and experience with multi-step computational scientific analyses.
• You must be able to validate scientific reasoning and computational outputs independently and create bounded, reproducible, objectively gradable research workflows.

Relevant experience includes bioinformatics, computational genomics, systems biology, computational neuroscience, biostatistics, computational drug discovery, computational biochemistry, structural biology, protein engineering, and computational microbiology.

• Scientific libraries and command-line tools.
• Reproducible research pipelines and computational life sciences publications.
• AI agents, coding agents, or automated evaluation environments.
• Docker, Linux, and domain-specific scientific software.

AI training is the human work behind systems that learn from examples, including scientific tasks, code, and model evaluations. People with specialized experience help check whether AI produces accurate, useful work in complex fields such as the life sciences.