Task Writer: Project Agentic-MME

$60/hr

rate as listed

Workload: ~15h/week
Engagement: Project-based
Location: Location not stated
Field: Creative

Create image-based evaluation tasks that require AI models to manipulate visual evidence, research facts online and verify the two together. Document the shortest valid solution path, authoritative sources, an exact golden answer and accepted variants.

The work

You will examine high-resolution images containing small, obscured or degraded details, test what frontier AI models miss and turn those failures into evaluation tasks. Each prompt must require meaningful image operations, such as cropping, rotation or enhancement, plus at least one web search.

For every task, you will document the shortest solution path step by step, identify the image operations used and cite a credible source for each search. You will also define one exact golden answer, the required answer format and any acceptable variants so the task can be scored consistently.

What Mercor screens for

The screening is likely to test whether you can isolate a genuine visual reasoning gap, design a realistic prompt and distinguish authoritative evidence from weak search results. Expect follow-up questions about how you would verify an image detail against a source, remove ambiguity from a rubric and decide which answer variants should receive credit. Advanced degree holders are sought, while prior rubric experience is a plus rather than a requirement.

Logistics

The listing does not specify whether the engagement is remote or asynchronous, nor does it state typical weekly hours or scheduling expectations. Confirm those details with Mercor before accepting work.

Pay band: $60/hr

What they look for

Advanced degree
Image-based task and prompt writing
Web research using primary, authoritative sources
Precise golden-answer and answer-format design
Step-by-step solution-path documentation
Rubric development or evaluation experience

Included with this listing

Interview prep — 10 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: This checks the stated degree requirement and whether the candidate can describe verifiable research methods rather than merely naming a qualification.
A strong answer: State the degree, institution and field, then connect a specific thesis, publication or research project to evidence gathering, source evaluation or precise written analysis.
Where people lose it: Naming the degree without explaining any concrete research process leaves the screener unable to test the claimed analytical background through follow-up.

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