Building interfaces
that convert.

Problem: Nepal's 30,000+ e-commerce vendors upload raw product photos — inconsistent backgrounds, wrong dimensions, poor lighting — directly to Daraz and Instagram. No affordable automated solution existed for the local market.

Solution: Full-stack AI pipeline that transforms raw vendor photos into catalog-ready images automatically. Vendors upload a photo, AI detects the product type, removes backgrounds with BRIA RMBG 2.0, reconstructs a photorealistic 3D studio presentation with GPT-image-1.5, and delivers 2000×2000 white-background images formatted for Daraz, SastoDeal, and Instagram — in under 60 seconds per image.

Problem: Nepal's premium mirror market had no dedicated online storefront — customers couldn't browse, compare, or inquire about custom mirrors without calling directly. Vendors managed orders manually with no admin visibility.

Solution: Full-stack e-commerce platform built from scratch for a Nepali mirror brand. Customers browse a product catalog, submit custom mirror inquiries, manage a wishlist, and check out with coupon support. Admins manage products, categories, orders, reviews, banners, and coupons through a custom dashboard — with Cloudinary image uploads and JWT-secured authentication throughout.