Introduction - davidmarsoni/llog GitHub Wiki

:bulb: General Information

Llog is a research project focused on integrating large language models with customized data sources for intelligent information retrieval and processing using agentic systems. The project is developed as part of the HES-SO Valais-Wallis 63-51 Emerging Technologies course.

:bookmark: Official sentence of the theme

Llamaindex: building context augmented AI assistants, is a flexible framework for building agentic generative AI applications that allow large language models to work with data in any format. (LLMs, agents, agentic Apps, RAG, generative AI) [CAJ]

:computer: Technology Stack

• Web framework: Flask, Tailwind CSS
• Query framework: LlamaIndex, Tavily
• LLM models: OpenAI
• Cloud storage: Google Cloud Storage
• Cloud deployment: Google Cloud Run

For detailed information about our technology choices and architecture, see the Technical Architecture page.

:dart: Project Goals

The main goal of this project is to understand the fundamentals of using LLMs and RAG approaches. To do this, we must be familiar with the LlamaIndex framework and the OpenAI API.

This project is primarily about prototyping and testing the integration of LLMs with customized data sources. The goal is to understand how to use LLMs and RAG approaches to build intelligent information retrieval and processing systems using agent systems.