AfterQuery is assembling a cohort of practicing mechanical and hardware engineers — CAD, RTL/HDL, and robotics — to write expert-level problems, reference solutions, and grading rubrics for frontier model evaluation. You judge whether a model's answer would survive a design review, not whether it sounds plausible.

Summary, eligibility guidance and screening preparation are prepared by Sidequest from the platform's posting. Check the source listing before applying.

What the work actually is

You pick a track — mechanical/CAD, RTL and digital hardware design, or robotics — and build evaluation material inside it. A typical unit of work is a self-contained scenario drawn from problems you've genuinely solved: a tolerance stack-up on an assembly that has to survive thermal cycling, a clock-domain-crossing bug that only shows up at a particular ratio, a manipulator whose Jacobian goes singular in the middle of a planned trajectory. You write the problem, you write the reference solution at the depth a senior engineer would defend it, and you write the rubric that says what separates a correct answer from an excellent one and from a confidently wrong one.

Then you grade. Model outputs in hardware domains fail in characteristic ways — plausible-looking free-body diagrams with a missing reaction, RTL that simulates but won't synthesize, control derivations that quietly assume a rigid-body model where compliance dominates. Your job is to catch that and to write the feedback explaining precisely which step went wrong and why it matters downstream. Structured written explanation is a substantial share of the hours, not an afterthought.

What the screen looks for

Currency of practice. They want people who have touched real parts, real silicon, or real robots recently. Purely academic-adjacent or long-lapsed backgrounds screen out.
Depth under follow-up. Expect the interviewer to push a second and third layer on whatever you claim — the specific standard, the specific failure mode, the number you'd actually use.
Rubric thinking. Whether you can convert your own tacit judgment into criteria another grader could apply consistently.
Writing. Technical explanation that stands on its own without a whiteboard.

Bonus signal comes from publications, open-source hardware or HDL contributions, and prior experience grading or reviewing others' technical work — peer review, design review, code review, teaching.

Logistics

Fully remote and asynchronous, structured as contract hours you schedule around existing employment. Observed rates run $100–170/hr depending on track and seniority; that band is what the platform advertises, not a guarantee for any individual engagement. Most contributors work in blocks of a few hours, with volume varying as cohorts and task batches open and close. There is a calibration period early on where your rubrics and gradings are compared against other experts'.

Mechanical / Hardware Design Engineer

What the work actually is

What the screen looks for

Logistics

Interview prep — 10 questions