Campus Second-hand Trading Platform

Full-Stack Development & LLM-Enhanced Search

2026.03 - 2026.04

Overview

A full-stack web application for campus second-hand trading, enhanced with LLM-powered intelligent search.

Database Design

Completed requirements analysis and ER modeling, designed a normalized database (BCNF) for users, products, orders, and reviews. Semi-automatically generated simulated data (50 users, 150 products, 100 orders) supporting the complete transaction process.

Front-end Implementation

Built a responsive web frontend with HTML/CSS/JS supporting:

  • User registration and login
  • Product publishing with image upload
  • Multi-dimensional filtering and search
  • Order status tracking
  • Review and reporting system
  • Administrator backend functions

LLM Intelligent Search Enhancement

Integrated the Qianwen API to leverage large language models for parsing natural language queries. When keyword matching fails, the system enables semantic recall and dynamic query relaxation, significantly improving product discovery capabilities.