Biostatistician

What you'll do

This is benchmark authoring. Each task you build simulates a real piece of biostatistics work: analysing a clinical trial, reviewing a regulatory submission, or assessing an observational study. You source and assemble the materials yourself (trial data, patient-level records, protocol documents), define the methodologically correct solution and a defensible clinical interpretation, then write a rubric of 35 or more items that an AI agent's attempt will be scored against.

The ad is explicit that textbook examples will not do. Tasks must carry the mess of real practice, which means the hard part is often finding or constructing data that is authentic enough to be a fair test while still having an answer you can defend line by line in a rubric.

Who fits

• An MS or PhD in biostatistics, statistics or epidemiology
• Four or more years in pharmaceutical, contract research, hospital or academic medical settings
• Command of clinical study design, regulatory documentation and observational methods
• Advanced R, Python, SAS or Stata on real clinical or epidemiological data
• Experience preparing FDA or EMA submission deliverables is a plus
• Comfort writing long, precise grading rubrics

What it pays

The listing shows $60–100/hour, but the compensation section says pay is output-based: you are paid per task that meets the project specification, and a weekly minimum number of submissions applies (the ad does not give the number). So the hourly band is a target, not a guarantee. A 35-item rubric plus a curated dataset is substantial work, and your first tasks will take longer than later ones.

Start timing

micro1 says it usually fills these roles within 48 hours and expects the first task within 24 to 48 hours of finishing onboarding. Apply only when you have capacity now.

Worth knowing

Good:

• Real regulatory and clinical experience is exactly what is being paid for
• Durable, intellectually serious work rather than rote grading
• 50 openings, posted recently and flagged high demand
• Fast start if you are selected

Less good:

• Per-task pay: rejected work is unpaid and no per-task fee is published
• A weekly minimum applies but is not quantified
• Sourcing authentic trial or patient-level data may be hard without access you already hold; the ad does not address data privacy
• Contractor engagement: no benefits, your own tax to manage

About this listing

Posted by micro1 on its own jobs portal nine days before we read it on 24 September 2026. The band, the 35+ item rubric size, the qualifications and the compensation terms come from that listing; no country restriction is stated. Payment schedule and visa terms are not published on the ad. See micro1.