Image Editing QA Reviewer

You will review edited images used to train AI image-editing models. Compare each edit with the original image, written prompt, and edit category to decide whether the requested change was made correctly.

• Check lighting, shadows, reflections, perspective, edges, color consistency, resolution, and visual artifacts.
• Confirm that identity, pose, background, and other details not mentioned in the prompt remain unchanged.
• Score each edit as approved, minor fix, or rework using a quality rubric.
• Write clear annotations and constructive correction guidance for image annotators.
• Flag unclear or inconsistent prompts, spot recurring errors, and contribute to guideline updates and scoring calibration.
• Keep reviews consistent while meeting daily accuracy and throughput targets.

What the role requires

• Workload: At least 20 hours per week.
• Location: Open worldwide.
• Language: Strong written English comprehension and the ability to explain visual decisions clearly.
• Relevant background in photo retouching, compositing, art direction, creative quality assurance, or similar work.
• A sharp eye for photorealism and subtle defects at full resolution, with the judgment to apply written guidelines objectively to unclear prompts.
• Working knowledge of Adobe Photoshop for checking edits and making small corrections.
• A desktop or laptop with a color-accurate display and reliable internet connection.
• Ability to work independently and collaborate with annotators and project leads.
• Helpful experience includes reviewing retouchers or designers, working in fashion, beauty, consumer goods, e-commerce, or retail studios, and doing data annotation or rubric-based quality checks.
• Familiarity with Photoshop Generative Fill, Firefly, Midjourney, or Stable Diffusion is also helpful.

How to apply

OpenTrain is a platform where people find AI training work, build a profile, and apply. Apply on OpenTrain with your resume, then complete the application on the hiring site.

About AI training work

AI training uses examples reviewed and prepared by people to help build and assess AI systems. Image reviewers use visual judgment to check whether edits follow instructions and preserve details that should not change.