AI Rater Guidelines Writer (Linguist / Instructional Designer)

What you'll do

Most AI training listings hire people to follow rater guidelines. This one hires the person who writes them. A leading AI lab needs linguists and instructional designers to turn loose program requirements into instructions that a pool of human raters can apply the same way every time.

The work, per the ad:

• Close the gap between ambiguous program requirements and rater-ready instructions across many subject domains
• Write clear, non-contradictory rating guidelines and rubrics that hold up under edge cases
• Audit draft guidelines for ambiguity, internal contradiction and coverage gaps, and revise until raters stop needing to escalate
• Translate specifications from finance, retail, insurance, legal and sports into precise, discipline-specific instructions
• Work with subject-matter experts and program leads to keep guideline sets consistent

Success shows up as fewer escalations, higher inter-rater agreement, and no two sections that point a rater in different directions on the same case.

Who fits

• 3+ years in linguistics, instructional design, technical writing or a close field
• Direct experience writing or refining guidelines or rubrics for human raters in a GenAI or RLHF setting
• Proven ability to work across several subject domains and make domain nuance unambiguous
• A track record of resolving ambiguity and contradiction in written specs, with concrete before-and-after examples you can show
• Demonstrable career progression
• At least 35 hours a week, reliably, on weekdays
• Strong, precise written communication

Prepare the before-and-after examples in advance. The ad asks for them specifically, and a single well-chosen revision (the original rule, the edge case that broke it, and your fix) will say more than a CV line.

What it pays

$45–65 per hour, Mercor's published band. That is below the Marketing Specialist and Retail Specialist listings posted on the same Cincinnatus template ($60–80), which also ask for 8+ years rather than 3.

How the engagement is structured

The body calls this a W-2 employment position with Cincinnatus LLC, placed with the lab's extended workforce; the header says part-time. The Mercor footer describes independent contracting with weekly Stripe or Wise payment. Ask which applies before accepting.

Worth knowing

Good:

• Writing the rules rather than following them, a step up from rating work
• Only 3 years of experience required
• Cross-domain variety: finance, retail, insurance, legal, sports
• Substantial weekly hours

Less good:

• US residents only
• Prior GenAI or RLHF guideline-writing experience is required
• 35+ weekday hours
• Conflicting W-2 and contractor terms
• H-1B and STEM OPT holders are excluded, per the Mercor footer

About this listing

Posted by Mercor as a part-time position with a US location requirement, confirmed open on 24 September 2026. Pay, hours and requirements are the ad's own. See Mercor.