The work
You would turn real agency and studio workflows into challenging evaluation tasks for generative AI models. Day-to-day work includes creating brand identity systems, packaging or out-of-home briefs, producing reference deliverables in editable source files, reviewing AI-generated work and documenting specific failures in typography, grids, hierarchy, color, consistency and production feasibility.
What Mercor screens for
The screening is likely to focus on verifiable shipped commercial work rather than speculative, student or crowd-rating projects. Be prepared to walk through a public portfolio, explain your contribution to identity and campaign systems, demonstrate hands-on fluency in Adobe Illustrator, Photoshop, Figma and InDesign, and discuss press-ready production using dielines, bleed, safe zones, CMYK, Pantone and PDF/X. It will also test whether you can distinguish a genuinely strong generative AI result from one that only appears polished at first glance.
Evaluation responsibilities
The role requires more than visual direction. You would build final files yourself, inspect editable source files and exchange formats such as .ai, .psd, .fig, .indd, .svg, PDF/X and dieline files, then develop rubrics and quality checks that research teams can apply consistently. Written feedback must identify observable design or production problems and provide enough reasoning to improve model behavior.
Practical details
This is a W-2 position with Cincinnatus LLC, with potential placement at a leading AI lab as part of its extended workforce. The commitment is at least 40 hours per week on weekdays; the listing does not specify whether the work is remote or how much of it is asynchronous.
Pay band: $70–120/hr
