What the work actually involves

You are a technical peer reviewer for long-horizon task materials set inside data-center and mission-critical projects. A typical unit of work hands you a scenario package — a bid set or issued-for-construction drawing extract, an EOR revision, vendor submittals, a procurement register, a 4D sequence, or a commissioning script — plus the prompt, the expected output, and the grading rubric. Your job is to find where it breaks: a spec section that contradicts the drawing, a lead time that makes the stated sequence impossible, a temporary-works assumption nobody stated, an acceptance criterion that no competent engineer would actually sign. Then you repair it — rewriting prompts, adding the missing constraint, tightening rubrics and programmatic checks — and validate the corrected task end to end.

The writing burden is real. Findings are read by people who were not in your head, so vague notes like "schedule seems off" do not land. What lands is "the 8-week switchgear lead time in the procurement register conflicts with the Week 6 energization milestone in the 4D sequence; either the milestone moves or a temporary power scheme has to be introduced, and the rubric currently rewards neither."

What the platform screens for

  • Delivery depth, not title depth. Screeners probe for projects you personally carried: scope, delivery model, your decision authority, what went wrong and what you did.
  • Interface literacy. Data centers fail at the seams — electrical to mechanical to structural to controls, EOR to trade contractor, submittal to installation. Expect follow-ups that push into a specific interface until you either show detail or run out.
  • Evaluation judgment. Can you tell an ambiguous task from a hard one? A task with two defensible answers is not automatically broken; a task with a rubric that only accepts one of them is.
  • Document-set fluency. Comfort reading specs, submittals, RFIs, schedules, and field records at volume without needing someone to summarize them.

Logistics

Remote and fully asynchronous. Approximately 5–10 hours per week initially, with an immediate start; volume can move with project demand. Pay is observed at $100–140/hr in this band, set by experience and role fit rather than guaranteed. PE, PMP, CCM, or commissioning credentials are preferred where relevant to your discipline but are not stated as required.