What the work involves

Projects vary, but they cluster around three tasks. The first is authoring: you take a development problem from a program you actually worked on — the modality, the process step, the specification it had to hit — and write it up alongside the reference answer a model's attempt will be graded against. The second is grading: reading model-written process development, assay development, or CMC-adjacent work and judging whether the scale-up assumptions survive contact with a real facility, whether a potency assay measures the mechanism it claims to measure, whether proposed release criteria are justified by anything other than convenience. The third is the subtle one: flagging proposals that are scientifically defensible and would still fail a developability assessment on manufacturability, stability, or what a regulator would accept.

Most of the output is written reasoning. A score with no explanation is close to worthless to the research team consuming it, so the writing is the deliverable as much as the judgment.

What the platform screens for

  • Whether you have taken a program through real development decisions, not just read about them. The screen probes for specifics: titers, hold times, comparability protocols, failed lots, the assay you had to requalify.
  • Whether you can reconstruct why a decision went the way it did, including the constraints that had nothing to do with the biology — timeline, facility fit, a regulator's prior feedback.
  • Written clarity under a word limit, since graders read hundreds of these.
  • Tolerance for ambiguous task instructions, and the instinct to flag an unclear rubric rather than guess and submit.

Where you trained matters less than the depth of programs you touched.

Logistics

This is a standing listing, not an open seat. Applying puts you in the biotechnology pool; you will not receive an accept or reject on this page. When a client project needs biotech expertise, matching people get invited to a specific listing that names the rate, the hours, and the client, and the hiring decision happens there. That invitation can come within a week or several months, depending entirely on what labs are running.

Work is remote and asynchronous, contract-based, with hours set per project. Observed rates in this field have run $110–125/hr, set per project by scope and depth — that is what past contracts have paid, not a guarantee. Application is a resume, a work-location confirmation, and an AI interview of roughly 20 minutes. Completing the domain expert interview once makes you eligible across every network you qualify for.