# Hybrid Restaurant Search System

Academic Project

A restaurant search system combining keyword search and vector search to find restaurants based on ambiguous user preferences.

[Home](/en-US) · [Projects](/en-US/projects) · [About](/en-US/about) · [Contact](/en-US/contact)

## Overview

A restaurant search system combining keyword search and vector search. It is designed to support not only structured search conditions but also ambiguous user preferences such as "a calm restaurant" or "a place suitable for going with friends."

## Background

Restaurant discovery often involves both explicit conditions and ambiguous preferences such as atmosphere, use case, and who the user is visiting with.

## Role

Frontend / Backend / Database / Deployment

## Tech Stack

- Next.js
- PostgreSQL
- pgvector
- MinIO

## Features

- Restaurant list views
- Keyword search
- Vector search
- Combined structured and semantic search
- Image and data management
- Web-service-ready architecture

## What I Focused On

I refined the overall system by combining and improving the outputs created by team members. I introduced frontend implementation, database integration, and MinIO, shaping the project into a web service that could be deployed publicly.


## Future Improvements

The current dataset includes temporary placeholder data. Future improvements include preparing real restaurant data, introducing the service to restaurants, and adding recommendation features.