Applied Mathematics Benchmark Specialist

$61–77/hrrate as listed

Write and verify graduate-level multiple-choice mathematics problems — each with nine subtly wrong distractors, a markdown chain-of-thought solution, and academic citations — for use as AI benchmark items. Work is asynchronous and remote, with an expected commitment of 10+ hours per week at an observed $61–77/hr.

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

The work

You will be assigned to one of two task streams. In authoring, you design original problems in an applied mathematics subdomain you actually work in, then build out the full item: one defensibly correct answer, nine plausible-but-wrong alternatives that a strong solver could fall for, a step-by-step chain-of-thought solution in markdown, a difficulty rating (Medium ≈ intro undergraduate, Hard ≈ advanced undergraduate, Expert ≈ post-graduate), and one to five references from peer-reviewed journals or university repositories. In verification, you receive pre-written items and audit them for correctness, ambiguity, missing assumptions, and solvability — editing where needed and documenting what you changed and why.

The hard part is rarely the mathematics. It is writing problems that are fully self-contained (no implicit conventions, no missing boundary conditions), and constructing distractors that correspond to specific, realistic errors — a dropped factor, a sign flip in a transform, an inverted risk-neutral measure, a misapplied Lagrange condition — rather than filler numbers. Reviewers push back on items where the correct answer is guessable by elimination or where two options are arguably both correct.

Domains in scope

Signal processing
Financial mathematics and actuarial science
Mathematical economics
Mathematical modeling of ecological and biological systems
Mathematical programming and combinatorial optimization
Geomathematics and climate modeling

What the screen looks for

Mercor's process is AI-led and resume-anchored: expect an interview that starts from your stated background and then probes one subdomain in depth, with follow-ups that test whether you can define terms precisely, state assumptions, and explain where a standard result fails. Credentials matter — PhD or doctoral candidate in mathematics, applied mathematics, or statistics, with master's-level candidates considered for exceptional subdomain depth — but the screen weights demonstrated precision and clear written English heavily. Prior competition writing, qualifying-exam authoring, or textbook problem-set development is a genuine differentiator.

Logistics

Fully remote and asynchronous; no fixed hours or standing meetings. Expected commitment is 10+ hours per week, and throughput depends on how quickly your items clear review — early submissions often take longer per item while you calibrate to the rubric. Observed pay is $61–77/hr; actual rate depends on domain, assessed depth, and task type, and is not guaranteed.

What they look for

PhD or doctoral candidacy in mathematics, applied mathematics, or statistics
Graduate-level depth in at least one listed applied subdomain (e.g., signal processing, financial mathematics, combinatorial optimization)
Formal proof writing and precise problem specification
Experience authoring rigorous assessment items, competition problems, or exam questions
Clear technical writing in English, including markdown and LaTeX-style notation
Ability to source and cite peer-reviewed literature

Before you apply

Would you pass this screen?

Drawn from what this listing actually requires. Answered in your browser — nothing is sent to us, and we do not store your answers.

Do you hold a PhD or are you a PhD candidate in Mathematics, Applied Mathematics, Statistics, or a closely related discipline—or do you hold a Master's degree paired with demonstrable research or professional expertise in one of the listed domains (Signal Processing, Financial Mathematics, Combinatorial Optimization, etc.)?
Advanced degree in mathematics or closely related field · required
Is English your native language, or do you have documented fluency (e.g., professional publication, peer-reviewed journal submission, teaching university-level coursework in English)?
Fluent English writing · required
Can you commit to a minimum of 10 hours per week on this work, on an ongoing basis, with a flexible schedule?
10+ hours weekly availability · required
Have you authored exam questions, written competition problems, designed tutoring materials, or reviewed academic assessments in a structured setting (not casual homework help)?
Experience in formal problem design or assessment · preferred

0 of 4 answered

Included with this listing

Interview prep — 11 questions

Matched to what this platform's screening actually measures. Answer each out loud in 60–90 seconds: claim → concrete example → implication.

What it tests: Whether the candidate understands the difference between a question that *feels* hard and one that tests deep conceptual mastery; ability to design plausible distractors that expose real misconceptions, not random noise.
A strong answer: A strong answer names a specific mathematical concept (e.g., confusing the direction of an inequality under transformation, or misapplying a theorem's preconditions), explains why a naive solver would fall into that trap, and shows how the wrong answers target different stages of flawed reasoning—not just "I made up wrong numbers."
Where people lose it: Candidates often describe difficulty as 'lots of steps' or 'obscure formulas' rather than conceptual vulnerability. Others admit their wrong answers are arbitrary; that signals they haven't authored rigorously or thought about student cognition.