Scientific Computing Expert: Life Science PhD ($300/Approved Task)

What you'll do

This role is aimed at a specific question: can an AI agent do real multi-step computational biology, not just answer a biology question? You design the tasks that test it. An agent is given scientific data and instructions and must pick methods, write and run code, notice when an intermediate result is wrong and fix it, then deliver a scientifically valid answer.

From the ad:

• Designing challenging agentic workflows that need several computational steps, not a single script
• Writing realistic input files, datasets, instructions, constraints and expected deliverables
• Building reproducible expert solutions with objectively verifiable ground truths and automated or semi-automated grading
• Validating assumptions, calculations, code and outputs so tasks are self-contained and run in network-isolated environments

This listing is for the Agentic Life Sciences project, unlike its Terminal Bench Science siblings.

Who fits

• PhD, postdoctoral or equivalent research experience in life sciences with strong computational work
• Strong scientific programming, particularly Python
• Experience with multi-step computational analyses and validating both the science and the code
• Expertise in at least one area: bioinformatics, computational genomics, systems biology, computational neuroscience, biostatistics, computational drug discovery, computational biochemistry, structural biology, protein engineering or computational microbiology

What it pays

$300 per approved task, as stated in the ad (and in its title). Tasks must pass review to count; the ad gives no quota.

Worth knowing

Good:

• Wide range of qualifying subfields
• Original design work rather than rote labelling

Less good:

• Pay depends on approval, so a hard-to-grade task can eat a day
• At least 6 hours a day and 40 hours a week, with 4 hours of Pacific Time overlap
• 5 weeks, start expected the following week
• Contractor assignment with no medical or paid leave
• The ad lists no assessment steps; Turing's help centre describes skill tests and sometimes live interviews

About this listing

Posted by Turing on 30 September 2026 and read on 18 October 2026. Rate and terms are Turing's own text; the comparison with sibling listings is ours. See Turing.