📖 About the Project
VibeVision AI was inspired by the emotional and highly unpredictable nature of crypto markets and the desire to convert raw data into clear, actionable signals.
While building it, I learned how to collect real-time financial data, engineer technical indicators, train and deploy neural networks, and connect AI systems with full-stack applications.
The project uses live data from Yahoo Finance, where indicators such as RSI, EMA Ratio, Bollinger Band Width, and normalized MACD form the feature set
[
X = [RSI, EMA\ Ratio, BB\ Width, MACD_{Norm}]
].
The model predicts whether the price will move up or down over the next 24 hours using deep learning.
Flask manages the backend, authentication, credits system, and prediction API, while React delivers a responsive cyberpunk-styled frontend.
Major challenges included handling noisy financial data, preventing model overfitting, synchronizing real-time inference with UI updates, and designing a fair credit-based economy.
Completing the system within hackathon time constraints required careful optimization and rapid iteration.
Overall, this project reflects the integration of AI, finance, and full-stack engineering into a single working product.
Log in or sign up for Devpost to join the conversation.