What the work actually is

You write chemistry problems hard enough that a frontier model gets them wrong, then supply the reference solution that proves it. That means original items — mechanism prediction, multi-step synthesis planning, spectral elucidation from raw NMR/IR/MS data, thermodynamic and kinetic reasoning, stoichiometry with real-world messiness — plus a worked answer whose every step survives scrutiny. A second stream of work is review: reading model outputs or other contributors' tasks and marking not just whether the final answer is right, but whether the reasoning that produced it holds. A correct product from an impossible mechanism is a failure, and saying why is the job.

One thing to know before you apply: the description Turing has posted under this Chemistry Expert title contains requirement text carried over from a legal listing (JD, bar admission, contract drafting). Treat the domain framing as chemistry — read the practice-experience, task-authoring, calibration, and escalation structure as the real shape of the role, and expect the screen to test chemistry depth. If a recruiter asks about bar admission, it is worth flagging the template error directly.

What the screen looks for

  • Depth that survives follow-up. The AI screener will push a second and third question into whatever you claim. Naming a pericyclic reaction is nothing; predicting stereochemical outcome and justifying it from orbital symmetry is the signal.
  • Specificity about your own bench or computational history. Which techniques you ran yourself, on what instruments, at what scale, and what went wrong.
  • Calibration instinct. Whether you can grade consistently against a rubric you did not write, and whether you escalate ambiguity instead of quietly inventing a house rule.
  • Honest edges. Analytical chemists who say they will not adjudicate organometallic catalysis score better than generalists who claim all of chemistry.

Logistics

Contractor engagement, fully remote, stated as up to 8 weeks. Minimum 4 hours per day, up to 40 hours per week, with at least 4 hours overlapping Pacific time — that overlap is the constraint that eliminates most non-US applicants who cannot shift their day. Pay is not disclosed in the listing; Turing typically sets hourly rates by domain and seniority at offer stage, so ask before onboarding rather than after. Shortlisted candidates receive a Job Interest Form before any next steps.