Job Finder Chatbot with RAG

This project is a chatbot application aimed at helping users find job opportunities and get relevant answers to questions about job roles. It leverages Retrieval-Augmented Generation (RAG) to provide personalized job recommendations and answers based on a user's skills, experience, and preferences. The system uses Azure OpenAI, Azure AI Search, and Azure PostgreSQL Vector Database for efficient and accurate search results. The backend is built with Java and Spring Boot.

Images

Features

• Job Search: Users can search for jobs based on their experience, skills, and preferences.
• Job Recommendations: The system suggests relevant jobs tailored to a user’s profile.
• Q&A System: Users can ask questions about job postings, such as required qualifications, experience, or job responsibilities, and get instant answers.
• RAG Integration: Retrieval-Augmented Generation is used to enhance the relevance of search results by combining real-time search and pre-trained language models.

Technology Stack

• Azure OpenAI: Used for generating embeddings and enhancing natural language understanding.
• Azure AI Search: Provides vector-based search capabilities to quickly retrieve relevant job postings based on embeddings.
• Azure PostgreSQL Vector Database: Stores job data with vector embeddings for efficient searching.
• Java: Core programming language for the backend.
• Spring Boot: Framework for building and managing the backend REST API.

Requirements

• Java 21 or later
• Maven for dependency management
• Azure account with the following services:
  • Azure OpenAI
  • Azure AI Search
  • Azure PostgreSQL
• PostgreSQL with the pgvector extension installed

Setup Instructions

1. Clone the Repository

Clone the project repository:

git clone https://github.com/shank250/CareerCanvas-msft-raghack

Navigate to the project directory:

cd job-finder-chatbot

2. Configure Database

• Set up Azure PostgreSQL and enable the pgvector extension for handling vector-based data.
• Update your PostgreSQL credentials in the application.properties or application.yml file in your Spring Boot project.

Example properties:

spring.datasource.url=jdbc:postgresql://:5432/ spring.datasource.username= spring.datasource.password=

3. Set Up Azure AI Search and OpenAI

1. Create an Azure OpenAI and Azure AI Search resource.
2. Add your API keys and Endpoint URLs to your Spring Boot application properties.

Example:

azure.openai.api-key= azure.search.endpoint= azure.search.api-key=

4. Run the Spring Boot Application

Build and run the Spring Boot application with:

mvn clean install mvn spring-boot:run

5. APIs

The application exposes the following APIs:

• /search-jobs: Allows users to search for jobs based on keywords, experience, and qualifications.
• /ask-question: Users can ask questions related to specific job postings and get instant responses.

6. Embedding Data

• During job posting ingestion, embeddings are computed using Azure OpenAI and stored in the vector column in Azure PostgreSQL.
• Queries are transformed into embeddings by Azure AI Search to retrieve the most relevant job postings.

Future Enhancements

• Support for multiple languages for job searches and queries.
• Expand the recommendation engine with more advanced ML models.
• Add user authentication for personalized job alerts.

License

This project is licensed under the MIT License.

Feel free to contribute and make improvements!