What the work actually involves
You write problems a model should be able to solve and then judge whether it did. In practice that means authoring case studies drawn from real production situations — a script breakdown that reveals a location conflict, a day that has to be re-boarded after weather, a budget that has to absorb an overage without cutting shoot days, a coverage plan for a two-hander with one camera. Each item needs a defensible correct answer and a written rationale explaining why the alternatives fail. The second half of the job is evaluation: reading model responses and marking where they are plausible-sounding but wrong — a schedule that ignores turnaround, a crew list missing a key position, a DI note that betrays no understanding of the offline-to-online handoff.
What the screen looks for
- Credits with specifics. Format, budget tier, your actual title, what you were responsible for. "Produced independent features" invites follow-ups you should welcome.
- Reasoning you can articulate. Many production decisions are tacit. The role requires converting them into prose a non-filmmaker grader could apply.
- Calibration. Recognising that several answers can be professionally acceptable, and being able to say what separates acceptable from wrong.
- Range. Someone who has worked across narrative, commercial, and live/theatre formats writes more useful items than a single-lane specialist.
Logistics
Fully remote and asynchronous — no shoot-day conflicts, no standing calls. Contributors typically pick up batches of item-writing or grading work and turn them around within a stated window; volume flexes, so this sits alongside active production work rather than replacing it. Expect a short calibration phase where your early items and gradings get reviewed before your throughput opens up. Rates in the $50–100/hr range have been observed on this listing and generally track domain depth and the difficulty tier you're trusted to author at.