About

• AI-Assistants Web Application includes two LLM-based applications using the OpenAI API.
• (1.) Language Assistant Tool, which translates and corrects texts in and from English, German and Hungarian.
• (2.) Asynchronous Real-Time Chat implemented with Django Channels and WebSocket.

Purpose

• I created this application as a user interface to use LLM-based tools online.
• It would be a kind of playground for trying use cases in several fields of expertise (natural language learning, education, decision support, data analysis etc.).
• The application is a basic API consumer now, using the web as an interface.
• The main goal is that the application can handle API communication between the user interface and fine-tuned and self-hosted LLMs.

Stack

• Backend: Django, Django-RestFramework, Django-Channels, PostgreSQL, Redis
• Frontend: ReactJS, CSS, Bootstrap
• Deployment: Docker, Railway.com

Development

• Trying and testing other APIs and models from another providers.
• Trying use cases from another fields of experties.
• Applying custom fine-tuned, self-hosted LLMs.
• Prompt engineering.
• Upgrading and trying new web designs.