What the work actually involves

You will be handed supply chain artifacts — a demand forecast and its assumptions, a sourcing scenario with competing supplier quotes, a distribution network tradeoff, an inventory policy with stated service levels — and asked to work them the way you would on the job, then judge whether a model's version of that work holds up. Some tasks ask you to author the problem and the defensible answer. Others ask you to grade two candidate responses and write out precisely where the weaker one breaks: a safety stock formula applied without regard to lead time variability, a landed cost comparison that quietly omits duty or demurrage, an EOQ recommendation that ignores MOQ constraints. The written rationale matters more than the verdict; a rating with no traceable reasoning is of little use to the people training on it.

Expect breadth. Projects rotate across manufacturing, retail, CPG, ecommerce and healthcare, and across the planning-to-execution spectrum. You may spend one week on S&OP consensus logic and the next on carrier selection and freight mode tradeoffs.

What the screen is looking for

  • Real operating experience, not coursework. One to three years actually running forecasts, POs, shipments or stock positions — the screener probes for specifics of systems used, volumes handled, and decisions you personally owned.
  • Numerical discipline. Whether you can carry units, cost components, and time buckets correctly through a calculation, and say what you'd need to know before answering.
  • Judgment about wrongness. Distinguishing an answer that is confidently wrong from one that is merely incomplete, and articulating the difference in writing.
  • Honest scoping. Naming the parts of supply chain you have not touched is treated as a signal of reliability, not a weakness.

Logistics

Fully remote and asynchronous — no standing meetings, no fixed hours. Projects are offered individually and you accept the ones you want; the current engagement runs roughly two to three weeks at 10–20 hours per week. Pay in the $50–100/hr range has been observed on this platform and varies by task complexity and reviewer track record; it is not guaranteed. Turnaround windows on individual batches are usually short, so consistent weekday availability matters more than total hours.