-p-1080.png)
1. Highly inconsistent product tags from sellers
Sellers uploaded items with missing, irrelevant, or incorrect tags, making the marketplace difficult to navigate and causing mismatched search results.
2. Large variation in product images
Images came in different lighting conditions, angles, backgrounds, and styles, requiring robust AI models capable of handling diverse inputs.
3. Manual review bottlenecks
Human reviewers spent significant time fixing tags and rejecting incorrect submissions, slowing down seller onboarding and delaying product go-live timelines.
4. Lack of a standardized taxonomy
Without a unified category and attribute structure, search and recommendation systems could not deliver accurate or relevant results.
