Skip to content

Latest commit

 

History

History
27 lines (19 loc) · 2.12 KB

readme.md

File metadata and controls

27 lines (19 loc) · 2.12 KB

🤖 Custom LLM Framework for Financial Analysis (Public Version)

Lucas Kemper – First Year MScFin Student at HEC Lausanne, Switzerland 🇨🇭

Inspired by my interests in quantitative finance 📊, AI 🧠, and LLMs 💬, I am customizing a fork of Lobe Chat to develop an innovative AI-powered framework for advanced financial analysis. Initiated on October 20, 2024, this project integrates state-of-the-art AI tools with finance-specific features designed to streamline data-driven analysis in investment and accounting.

Note: The code remains private to maintain confidentiality as I refine and enhance this framework. 🔒

🌟 Core Features

  1. File Upload & Knowledge Base 📁
    Personal knowledge repository enabling efficient file upload and search, tailored for financial datasets (upload functionality currently under optimization).

  2. Multi-Model Support 🤝
    Integrated support for OpenAI and Anthropic models to provide versatile AI options.

  3. Advanced Infrastructure Optimizations

    • Vercel: Front-end deployment optimized for speed and cost efficiency; transitioning to AWS for enhanced performance and security (Status: In Progress) 🚀
    • Cloudflare: Added comprehensive security with DNS, CDN, and DDoS protection (Status: ✅ Completed)
    • MongoDB to AWS Migration: Migrated database from MongoDB to AWS PostgreSQL (using Prisma) for enhanced data management efficiency (Status: ✅ Completed)
    • Clerk Authentication: Implemented secure, multi-provider login through GitHub and Google (Status: ✅ Completed)

🎯 Current Challenges

  • Model Complexity Constraints: Addressing limitations in context length for complex financial queries (currently experimenting with Hugging Face) 🔄
  • File Upload Bug: Finalizing network configurations for seamless file processing (Status: ✅ Completed)
  • API Cost Optimization: Focusing on cost-efficient API usage 💰
  • Data Quality: Enhancing data input accuracy, essential for quality financial analysis 📊 (Status: ✅ Completed)