I Gave a Brain to an Online Store

It’s no longer just about putting items in a cart. Shoppers expect quick answers, real personalization, and a smooth experience from start to finish.

Building my final year project reminded me how much online shopping has changed over the years.

It’s no longer just about putting items in a cart. Shoppers expect quick answers, real personalization, and a smooth experience from start to finish.


That’s where GizmoGalaxy comes in, my final year project, developed to close the gap between what users expect from online shopping and what most platforms deliver.

I wanted to create something intelligent. A system that understands what the user needs, recommends what makes sense to them, and even chats with them when they’re stuck.

So, I built:
🧠 A smart assistant(with gemini-2.5-flash) that chats, guides, and helps you shop
🎯 Behavior-based product recommendations tailored to each shopper
🔍 Voice + semantic search that actually understands natural language
🛒 A complete admin dashboard to manage everything behind the scenes

The tech stack includes Next.js, Express.js, PostgreSQL (via Supabase), and Clerk for authentication, with deployments on Vercel (frontend) and Azure App Service (backend).

But here's the important part, this isn’t just built for showcasing tech items.That’s just the prototype.
What I’ve built is a flexible, domain-agnostic MVP that can easily be adapted for any kind of e-commerce business.

📦 launching a clothing brand?
🍔 Running a food store?
🪑 Selling furniture?

This system is already designed with that in mind, and I’m currently working on adapting the core for different verticals with custom flows and feature sets.

If you’re into AI, user experience, or just building better online stores, I’d love to hear your thoughts.
Finally, a huge thank you to Dr. Mohamed Shafraz , Ph.D. my supervisor, for his continuous support and guidance throughout this project.


Up next