Image Nimbus
AI-assisted product image scraper and sourcing system
Shifted the work from finding every image by hand to reviewing ranked candidates and uncertain matches, which made image prep easier to keep moving.
Built a Python scraper and sourcing system that pulled product image candidates at scale, then ranked matches so the team could review the best options instead of starting every search from scratch.
The Problem
Team members were manually searching for product images one by one, which meant slow turnaround times and a lot of products with missing images. The volume of products made it impossible to keep up.
What I Built
Built a scraper-driven workflow with AI evaluation that scores each image against the product details: first checking if it's the right category, then brand, style, and color. Offshore workers only review the lowest-confidence matches where the system isn't sure, so the team can focus on quality control instead of hunting images.
Who It Helped
This helped listing operations and offshore reviewers start from ranked image matches instead of blank manual searches.
