What the work involves

This is a two-track assignment. On the data generation side, you construct the pharmacology layer of simulated clinical data rooms: concentration-time profiles, dose and escalation schedules, PD biomarker trajectories, exposure-response relationships, and bioanalytical method and results reports. The constraint that makes it hard is internal consistency — the exposure data has to reconcile with the efficacy readouts and the safety signals in the same fictional program, so an expert reviewer reading the whole package can't find the seam. Working from a protocol and SAP to produce numbers that hold up under scrutiny is the core skill.

On the authorship side, you write the artifacts a clinical pharmacologist would produce — dose selection rationale, drug concentration and response sections of a clinical study report, DDI and drug-disease interaction assessments — and then convert each into an evaluation item: a prompt an AI agent receives, a golden output representing what a competent pharmacologist would produce, and a rubric that separates a defensible answer from a plausible-sounding wrong one. Rubric design is where most of the intellectual weight sits, because the failure modes you're trying to catch are subtle: an agent that cites the right PK parameter with the wrong units, or builds a dose rationale on a safety margin it never actually computed.

What the platform screens for

Mercor's screen is AI-led and pushes on specifics. Expect follow-ups on phase-specific work: what you did on a first-in-human dose escalation, how you chose a starting dose, how you handled an exposure-response analysis that didn't support the sponsor's preferred dose. Generic "I've worked in clinical pharmacology" answers get probed until they either produce detail or collapse. Pre-market Phase 1–3 experience as a clinical pharmacologist is a stated hard gate, not a preference — post-marketing, preclinical-only, or purely academic PK modeling backgrounds are typically screened out here.

Logistics

  • Remote, US or Canada only.
  • 20–25 hours/week minimum; 30+ preferred. This is not a few-hours-on-the-weekend engagement.
  • Largely async, with review cycles against other contributors building the efficacy and safety layers of the same data rooms.
  • Observed pay band $120–170/hr, set by experience and depth; bands on this platform are observed, not guaranteed.