Lunova - Modern Jewelry & Souvenir E-commerce Platform

Voice Assistant
Dashboard #1
Dashboard #2
Landing
Architecture

🌟 About the Project: Lunova App

Lunova was born out of a vision to redefine the premium jewelry and souvenir shopping experience using the power of AI, serverless computing , and micro-frontend architecture. In 1 week, I set out to build a modern, scalable, and intelligent platform that not only looks great but also empowers both users and shop owners with smart tools, automation, and global reach.

🚀 What Inspired Me

The luxury gift and jewelry market often lags behind in digital innovation. Most platforms still rely on static catalogs and generic recommendations. I envisioned Lunova as a smart virtual shop that uses AI and real-time insights to make premium shopping elegant, intelligent, and personalized. The idea was simple: combine the power of generative AI, vector search, and modular design to enhance every step of the shopping journey.

🛠️ What I Built

Lunova was engineered using a modern serverless stack that prioritizes performance and developer speed:

✅ Frontend with React, Vite, and GSAP animations — deployed via Netlify Functions
✅ Authentication with Supabase Auth (JWT-based, fully serverless)
✅ Database & Storage: Supabase PostgreSQL + Supabase Storage
✅ Real-time updates using Supabase Realtime for notifications and interactions
✅ Edge Functions (Supabase) for API logic + FastAPI on serverless containers for high-performance routes
✅ OpenAI + AWS Lambda to build a virtual shop assistant (LLM-based chatbot)
✅ Qdrant Cloud Vector DB for similarity search and product embedding storage
✅ Perplexity API + AWS Lambda to deliver global market insights for shop owners
✅ AWS Bedrock (Claude) for customer support ticket generation
✅ CRM Dashboard powered by FastAPI + Supabase
✅ Clean schema design with dbdiagram.io

📚 What I Learned

Mastered building cloud-native applications with a 100% serverless backend
Integrated LLMs and LangChain with vector databases for contextual product discovery
Built a robust, modular architecture combining Supabase, Lambda, and FastAPI
Improved UX through real-time updates, smart suggestions, and rich frontend interactions
Tackled multi-language challenges across content and user interfaces

🧗 Challenges I Faced

Coordinating multi-service orchestration between Supabase, AWS, OpenAI, and Qdrant
Managing latency and cold starts in Lambda functions
Building a seamless flow for image uploads, metadata tagging, and vector embedding
Maintaining consistency across languages (English, Japanese) and dashboards
Scaling the chatbot with meaningful and personalized product recommendations

💡 Why Serverless Made It Possible

Lunova wouldn't be possible without the serverless paradigm. It allowed me to:

🚀 Ship faster without managing any infrastructure
📉 Keep costs low during development and scale effortlessly as traffic grows
🌎 Deploy globally with edge functions for low latency worldwide
🔁 Build modular features (auth, insights, AI assistant, chat) as independent, serverless units
🎯 Focus entirely on product logic and UX - not on infrastructure

Lunova represents more than a hackathon project - it’s a startup-ready foundation for a smarter, global, and AI-enhanced e-commerce experience. ✨