What the work actually looks like

You will be handed model-generated answers to CFD questions and asked to decide whether they are correct, and if not, exactly where they break. In practice that means things like: a model proposes a k-ω SST setup for a separated diffuser flow and you have to judge the wall treatment and y+ target; a model writes an OpenFOAM `fvSchemes` block and you check whether the divergence scheme is stable for the Reynolds number in question; a model reports a drag coefficient that looks plausible but comes from an unconverged run. Other batches ask you to author problems in the first place — a well-posed case with boundary conditions, a reference result, and a rubric another reviewer could apply without guessing.

Most tasks are self-contained and take 20–90 minutes. Some require you to actually run something locally; many do not, and are judged on physics and numerics reasoning alone. Expect written rationales, not just scores — the rationale is the product, because it is what teaches the model.

What the screen is looking for

micro1's screening is AI-led and heavy on follow-up. It probes whether your CFD knowledge is operational or vocabulary-deep: you will be asked about turbulence closure choices, mesh independence, discretisation error, convergence criteria, and boundary condition selection, and then asked why immediately afterward. Vague appeals to "best practice" collapse under one follow-up. The other thing it screens for is calibration — whether you can say a model answer is mostly right with one fatal flaw rather than defaulting to pass or fail.

  • Verifiable background: named solvers, named projects, degree level, industry or research setting
  • Ability to write a defect description precise enough that an engineer could fix the case from your text alone
  • Comfort reading and critiquing code, not just GUIs

Logistics

Fully remote, asynchronous, contract. Work arrives in batches and volume is genuinely uneven — some weeks nothing, some weeks a queue. Most contributors treat this as 5–20 hrs/week alongside other work. Rates are set per project and per assessed level; the $50–150/hr band reflects what has been observed on similar micro1 engagements rather than a guaranteed offer. Payment is per approved task or per hour depending on the project's structure.