RiskXpertAI

Project Duration: Mar 2025 – Apr 2025

AI-Powered Financial Indicator Forecasting Chatbot

• Tech Stack: Docker, GitHub Actions, AWS, FastAPI, MongoDB, Redis, Kakao Login API
• Team: 2 members (My Role – Full service development & operations, Teammate – AI modeling)
• GitHub: RiskXpertAI
• Deployment: RiskXpertAI

📌 Project Overview
• A personalized financial forecasting service that supports investment strategy and responds to market volatility.
• Leverages open-source AI models and real-time data processing to deliver concise and rapid insights.

🎯 Motivation & Goals
• A need for real-time, personalized tools to understand and react to financial market trends.
• Build a forecasting and analytics platform integrating high-performance AI models based on open-source.

🚀 Key Features
• Real-time response with streaming-based processing
• Fast response through Redis caching and MongoDB dialogue storage
• Time series prediction using Transformer-based models
• Scenario summarization and financial report generation using GPT-3.5
• Integrated with Kakao OAuth authentication

💻 Responsibilities & Contributions
• Developed FastAPI-based backend and inference APIs
• Designed complete infrastructure architecture (Redis, MongoDB, Docker, etc.)
• Implemented real-time response flow and Slack error alert system
• Built CI/CD pipeline with AWS EC2, Docker, GitHub Actions for zero-downtime deployment

✅ Outcomes & Reflection
• Operated a lightweight architecture optimized for low traffic
• Built a practical AI service balancing cost efficiency and system stability