Resume Chatbot

Overview

LLM Chatbot(ResumeBuddy) built using Supabase for vectoreStore, OpenAI API for embeddings with adjusted chunk size and overlaps for Resumes.

Features

• Natural Language Understanding: Leverages OpenAI's GPT for understanding and responding to user queries.
• Document Embeddings: Utilizes vector embeddings for efficient document retrieval.
• Dynamic Conversation Handling: Maintains and uses conversation history for context-aware responses.
• Scalable Architecture: Built with Express.js and Supabase, ensuring scalability and performance.
• User-Friendly Interface: Simple and intuitive UI for seamless interaction.

Technologies Used

• Frontend: HTML, CSS, JavaScript
• Backend: Node.js, Express.js
• Database: Supabase
• AI/ML: OpenAI GPT, LangChain
• PDF Parsing: pdf-parse
• Environment Management: dotenv
• Build Tools: Vite

Installation

1. Clone the repository:

   git clone https://github.com/yourusername/resumebuddy.git

2. Navigate to the project directory:

   cd resumebuddy

3. Install the dependencies:

   npm install

4. Set up environment variables: Create a .env file in the root directory and add your API keys and other environment variables:

   OPENAI_API_KEY=your_openai_api_key
   SUPABASE_API_KEY=your_supabase_api_key
   SUPABASE_URL_LC_CHATBOT=your_supabase_url
   PORT=3000
   

Usage

1. Start the server:

   npm start

2. Open your browser and navigate to http://localhost:3000 to interact with ResumeBuddy.

Project Structure

• index.html: The main HTML file that serves the chatbot UI.
• server.js: The main server file that handles API requests and interactions with the AI model.
• package.json: Contains project metadata and dependencies.
• txt_embed.js: Script for embedding text documents into the vector store.
• combineDocuments.js: Utility to combine multiple documents into a single text.
• formatConvHistory.js: Utility to format conversation history.
• retriever.js: Script to set up the document retriever using Supabase and OpenAI embeddings.

How It Works

1. PDF Loading and Splitting: The resume PDF is loaded and split into manageable chunks using RecursiveCharacterTextSplitter.
2. Embedding and Storing: The chunks are embedded into vector representations and stored in a Supabase vector store.
3. Question Handling: User queries are processed, and standalone questions are generated if necessary.
4. Context Retrieval: Relevant document chunks are retrieved based on the query context.
5. Answer Generation: The AI model generates context-aware answers using the retrieved information.

Demonstration

Home Page

Personal Projects Query

Memory Retention - Initial Query

Memory Retention - Follow-up Query

Terminal Output Demonstrating Memory Retention

Contribution

Feel free to open issues or submit pull requests for improvements and bug fixes.