ResumeGPT - Art Kreimer's AI-Powered Resume Chatbot

A modern, interactive chatbot application that answers questions about Art Kreimer's professional background, built with React, TypeScript, Vite, and Supabase.

Overview

This application uses RAG (Retrieval Augmented Generation) with OpenAI's GPT-4 to provide accurate, context-aware responses about professional experience, skills, and qualifications. The chatbot is deployed as a serverless application using Supabase Edge Functions and can be hosted on Vercel.

Features

• Modern React UI: Clean, responsive design built with React, TypeScript, and Tailwind CSS
• Real-time Chat: Interactive chat interface with suggested follow-up questions
• AI-Powered Responses: Uses OpenAI GPT-4 for intelligent, contextual answers
• Conversation Storage: All conversations are stored in Supabase for analytics
• Serverless Architecture: Edge Functions handle chat logic, scalable and cost-effective
• Easy Deployment: Ready to deploy on Vercel with zero configuration

Tech Stack

• Frontend: React 18 + TypeScript + Vite
• Styling: Tailwind CSS
• Backend: Supabase Edge Functions
• Database: Supabase PostgreSQL
• AI: OpenAI GPT-4
• Deployment: Vercel

Prerequisites

• Node.js 18 or higher
• A Supabase account with a project
• OpenAI API key
• Vercel account (for deployment)

Local Development

1. Clone the repository

git clone https://github.com/yourusername/resumegpt.git
cd resumegpt

2. Install dependencies

npm install

3. Set up environment variables

Copy .env.example to .env and fill in your values:

cp .env.example .env

Update the .env file with your Supabase credentials:

VITE_SUPABASE_URL=your_supabase_project_url
VITE_SUPABASE_ANON_KEY=your_supabase_anon_key

Customization

To customize this chatbot for your own resume:

1. Update the knowledge base in supabase/functions/chat/index.ts:

   • Modify the knowledgeBase array with your own Q&A pairs
   • Update references to "Art Kreimer" with your name

2. Update the system prompt in the same file:

   • Change the chatbot's personality and instructions
   • Modify the response format if needed

3. Customize the UI in src/App.tsx:

   • Update the welcome message
   • Change the header title
   • Modify the default suggested questions

4. Update branding:

   • Change the title in index.html
   • Update the favicon
   • Modify colors in tailwind.config.js or component styles

Deployment

Deploy to Vercel

1. Push your code to GitHub

2. Import your repository to Vercel:

   • Go to vercel.com
   • Click "Add New Project"
   • Import your GitHub repository

3. Configure environment variables in Vercel:

   • VITE_SUPABASE_URL: Your Supabase project URL
   • VITE_SUPABASE_ANON_KEY: Your Supabase anon key

4. Deploy: Vercel will automatically build and deploy your application

Project Structure

resumegpt/
├── src/
│   ├── components/
│   │   ├── ChatInput.tsx       # Message input component
│   │   ├── ChatMessage.tsx     # Individual message display
│   │   └── SuggestedQuestions.tsx  # Suggested questions UI
│   ├── lib/
│   │   └── supabase.ts         # Supabase client configuration
│   ├── App.tsx                 # Main application component
│   ├── main.tsx                # Application entry point
│   ├── types.ts                # TypeScript type definitions
│   └── index.css               # Global styles with Tailwind
├── supabase/
│   └── functions/
│       └── chat/
│           └── index.ts        # Edge Function for chat logic
├── data/                       # Original data files (reference)
├── index.html                  # HTML entry point
├── vercel.json                 # Vercel configuration
├── package.json                # Dependencies and scripts
├── tailwind.config.js          # Tailwind CSS configuration
├── tsconfig.json               # TypeScript configuration
└── vite.config.ts              # Vite configuration

Database Schema

The application uses a single table for storing conversations:

conversations

• id (uuid, primary key)
• conversation_id (uuid) - Groups messages from same session
• user_message (text) - User's question
• bot_message (text) - Bot's response
• answered (boolean) - Whether question was answered successfully
• created_at (timestamptz) - Timestamp

Acknowledgements

This project is highly influence by the Repository created by Art Kreimer and is dependent on the following libraries and tools:

• Streamlit for building the web application
• LangChain for integrating the language model and retrieval chain
• OpenAI API for the language model
• FAISS for the vector database
• Firebase for storing the conversation history

License

This project is licensed under the MIT License.