The work
You author interpretive scenarios — a passage, a prompt, a question that has a defensible answer grounded in the text — and you write the reference reading that goes with it. The reference reading is the real deliverable: not a summary of what the poem is about, but an argument with the specific lines, syntactic moves, formal choices, or intertextual signals that license each claim. Graders and models are both checked against it, so vagueness costs you.
The second half of the job is evaluation. You read model output and score it for textual fidelity (does the claim survive contact with the passage?), interpretive depth (does it do anything beyond restate theme?), and craft awareness (does it notice enjambment, free indirect discourse, register shift, an unreliable frame?). A recurring task is flagging readings that sound like competent criticism but float free of the text — the model that confidently attributes an image to a stanza that doesn't contain it, or that hangs a thesis on a word choice the author never made.
What the screen looks for
AfterQuery's screening is AI-led and follow-up heavy. It probes whether your close reading is real: expect to be asked to name a specific text, quote or paraphrase a specific line, and say what the evidence rules out as well as what it supports. It also tests calibration — whether you can distinguish an interpretation you disagree with from one that is wrong on the text, since collapsing those two is the single most common failure among strong readers new to evaluation work. Have your degree, your teaching or editorial history, and any publications ready in verifiable form.
Logistics
- Fully remote and asynchronous; no fixed meetings, no shift coverage
- Minimum commitment of 10 hours per week, scalable up or down
- Weekly payment via Stripe; the rate within the $50–100 band is set from experience and is observed, not guaranteed
- Assignments are routed toward your period, genre, or tradition where volume allows