AI Image Quality Auditor & Creative Advisor (E-commerce)

About the role We've built an internal tool that automatically generates creative product imagery for e-commerce listings (Amazon, Walmart, Target) — style guides, creative briefs, and finished creative images produced by an AI pipeline. Our client success managers run it day-to-day. What we're looking for is an expert eye above them: someone to audit the output, hold a quality bar, and advise on the technical fixes when the AI gets it wrong. This is an auditor + advisor role, not a production seat. You review and direct; our team executes. What you'll do - Audit AI-generated creative in batches and grade it against a professional standard: composition, hierarchy, typography, color, brand consistency, product fidelity (does the render still look like the real product?), and commercial polish. - Help define and maintain what "ship-ready" looks like, and flag what falls short — and why — in clear, actionable written feedback. - Consult on the technical fixes. When an image comes out wrong (warped product, off-brand palette, bad text rendering, weird "before/after" artifacts), diagnose why at the model/prompt level and prescribe the fix: prompt adjustments, reference-image strategy, model choice, parameters, or editing/inpainting approach. - Partner with our team to translate design problems into technical direction they can implement. You're a strong fit if you have - A professional graphic design background with a portfolio (e-commerce, product, or advertising work strongly preferred) — enough authority to be the one grading the work. - Hands-on experience with modern AI image tools (e.g. Midjourney, DALL-E / GPT image, Stable Diffusion, Flux, Nano Banana / Gemini image, ComfyUI, inpainting/editing, ControlNet or reference-conditioning). - The ability to articulate why an AI image fails and how to steer it — prompt engineering, reference images, seeds, model selection, post-editing. - Strong written communication; your deliverable is critique and direction, so it has to be precise. - Bonus: familiarity with Amazon listing imagery standards, Photoshop/Figma, or light scripting. Not a fit if you only do hand-crafted design and treat AI tools as a black box, or you can generate AI images but can't critique them like a professional designer. We need both — the eye and the technical fluency. Engagement Ongoing, part-time / as-needed advisory. Remote and async-friendly. We'll start with a small paid trial: we share a set of generated outputs, you audit them, deliver written critique against a quality bar, and prescribe technical fixes for the weakest ones. How to apply Send a short Loom video (3-5 min) in which you: 1. Introduce yourself and your design background (share portfolio links). 2. Critique one AI-generated image — walk us through how you assess its quality. 3. Show one AI image that came out wrong and explain how you'd diagnose and fix it at the model/prompt level. 4. Tell us which AI image models/tools you know best and why. Applications that include a Loom video will be prioritized.