BACKGROUND

This is my playground for various LLM and AI experiments. Initially it will be a resume question-answering chatbot, maybe adding the ability to convincingly interview for a job through A2A later. Eventually I'd like to vectorize every e-mail, chat, essay, social media post, etc. I've archived and use RAG or perhaps model fine-tuning to create a more complete "Virtual Ryan".

The current implementation is using GitHub Models to provide speed and a large context window for all the background information in my resume and documentation.

Thanks to the UI geniuses whose work I pilfered to give my project some polish.

• The Max Headroom-esque background: https://codepen.io/0x04/pen/DoEELw by Oliver Kühn
• Chat interface: https://codepen.io/supah/pen/jqOBqp by Fabio Ottaviani

INSTALLATION

1. Clone the repository:

   git clone https://github.com/rrusson/VirtualRyan.git

2. Install dependencies:

   npm install

3. Set up a GITHUB_TOKEN for your account, if you don't have one.
   • Go to https://github.com/marketplace/models/azure-openai/gpt-4-1 and click the green "Use this model" button to get started.
   • The token must include the models:read permission for a fine-grained Personal Access Token or use a Classic token with repo scope.
   • More information about tokens here.
4. Set a system environment variable for your GITHUB_TOKEN (instructions for Windows 10)
5. Follow the customization guide to personalize the application