A full-stack, AI-powered Lost & Found platform built for Discovery High School (US) to modernize and automate item recovery workflows using AI, real-time coordination, and secure verification systems.

This project replaces the traditional manual lost-and-found process (physical bins, social media posts) with a centralized digital system that enables:

• Intelligent AI-based item matching
• Secure multi-user recovery workflow
• Automated email-based notifications
• Admin analytics & monitoring dashboard

This is designed as a real-world engineering system, not a CRUD application.

The current system at Discovery High School suffers from:

• Low recovery rate of lost items
• No centralized tracking database
• Manual, inefficient claim process
• Lack of verification between users
• No analytics for administrators

This project solves these issues through automation, AI, and structured workflows.

• JWT-based authentication system
• Email verification required for account activation
• Role-based access control (Student / Admin)
• Two-factor authentication using image-based login
• Fraud prevention using unique pickup verification codes

• OpenAI CLIP image embeddings
• Cosine similarity-based image matching
• Hybrid search (image + metadata filtering)
• Ranked similarity scoring with percentage output
• Personalized AI suggestions (owner-only visibility)

• Email-driven event notification pipeline
• Triggers for:
  • New found item postings
  • AI-matched item detection
  • Admin announcements
  • Pickup status updates
• Supports both image-based and text-based matching alerts

• Multi-step recovery process:
  • “I will pick up” request
  • Schedule negotiation between users
  • Confirmation of meeting time
  • Verification code exchange during handoff
• Dynamic rescheduling system
• Email synchronization for every workflow state change

• Bar chart: Lost / Found / Returned items
• Pie chart: system distribution overview
• Line chart: inventory trend over time
• Tracks:
  • Recovery rate
  • Item backlog
  • System efficiency metrics

• Voice-controlled navigation using Web Speech API
• Hands-free interaction support
• Accessibility-first UI design

Frontend (React + Vite)
↓
REST API (ASP.NET Core .NET 8)
↓
Service / Business Layer
↓
AI Matching Engine (CLIP + Cosine Similarity)
↓
Database (SQL Server)
↓
Email Notification System (SMTP)

• Hosted on Microsoft Azure
• Optimized for:
  • High concurrency handling
  • Async email processing
  • Scalable API architecture

• React.js (Vite)
• Web Speech API

• ASP.NET Core (.NET 8)
• SignalR (real-time updates)
• JWT Authentication

• OpenAI CLIP embeddings
• Cosine similarity engine

• SQL Server

• Microsoft Azure

• Designed a multi-user state-driven workflow system
• Built AI-powered multimodal search engine
• Implemented event-driven email notification architecture
• Solved real-world constraint: no push notifications → email-based system design
• Created fraud-resistant verification system
• Integrated AI + workflow + security + analytics into a unified platform

• Improved item recovery efficiency through AI matching
• Reduced manual workload for school staff
• Increased transparency between users
• Enabled secure real-world item handoff
• Provided actionable analytics for administrators

• WebSocket-based push notification system
• Vector database integration (FAISS / Pinecone)
• Location-based matching system
• AI ranking model optimization
• Mobile-first redesign