What the work involves
You author scenarios that look like real casework: a self-referred client disclosing passive ideation midway through an intake, a caregiver navigating discharge planning with no insurance, a mandated reporting judgment call where the facts are genuinely ambiguous. For each one you write a reference response and — this is the part labs actually pay for — the clinical reasoning behind every decision in it: why you asked that question before this one, why you assessed means before intent, why you did not offer a resource list at that moment.
The second half of the job is grading. You read model output against your scenarios and score it for clinical accuracy, ethical handling, and risk management. Much of what you flag will be output that reads warm and validating but would fail an actual client — the response that reflects feeling and never screens for safety, or that hands a client a hotline number as a substitute for assessment. You have to name the failure precisely enough that a non-clinician engineer can act on it.
What the platform screens for
AfterQuery verifies the LCSW license and post-licensure practice years first; those are gates, not preferences. Beyond that, screening probes whether you can articulate clinical reasoning rather than just perform it — expect follow-ups that push on why, not what. Expect at least one live scenario where you're shown a plausible-sounding response and asked to say what's wrong with it. Written clarity is assessed from your actual answers, since the deliverable is written documentation.
Logistics
- Fully remote, fully async — no scheduled shifts, no client contact, no caseload
- Minimum 10 hours per week; you set the schedule inside that
- Paid hourly via Stripe, weekly, at an observed $70–105/hr band that varies with specialty and task type
- Ongoing project work rather than a fixed-term engagement