Duration: 4 Weeks | Start Date: Immediate | Location: UK Only

Rate: £16 per hour (£120 per day based on 7.5 hours)

The Role

This is a technical visual role for someone with an obsessive eye for detail. We are migrating our furniture database into new retail partner formats. You will be working with a library of 100,000+ images, responsible for pulling the correct sofa models, fabric variants, and ensuring all 7 angles per variant are present and correct.

This is not a simple upload task. You will be running renaming scripts, batch-processing assets, and, most importantly, manually auditing the quality of the renders to ensure they accurately represent the product.

Requirements

Responsibilities:

  • Asset Extraction: Navigate a database of 100k assets to pull specific SKU variants.
  • Format & Resize: Adapt images to strict retailer specifications (aspect ratio, padding, file type).
  • Renaming & Scripting: Run and monitor renaming scripts; troubleshoot any logic breaks in the file-naming convention.
  • Visual QA: Identify inconsistent renders where lighting, texture, or colour profiles do not match the master fabric swatches.
  • Retouching & Upload: Perform light retouching where necessary and manage bulk uploads to retail partner platforms.

The Ideal Candidate:

  • Technical Precision: You understand the difference between a "glitch" and a "mapping error."
  • Visual Sensitivity: You can immediately spot when a 3D render doesn't accurately reflect the selected fabric swatch.
  • Efficiency: Able to maintain high output across 4 weeks without losing focus on the small details.

How to Apply

To filter for the high level of attention to detail required, we will only review applications that follow these steps in their covering letter:

Subject Line: Must be exactly: "Visual Audit: [Your Name]"

The Logic Test: Visit this product page and select the fabric "Pillarbox Easy Velvet".

  1. Compare the Colour Selector (the swatch) to the Product Images in the carousel. What is fundamentally wrong with the render of the sofa?
  2. In a project involving 100,000 images, why is this specific type of error more problematic than a missing image?