QuantumLeap ❇️

AlphaStream Engine addresses the complexity of finding investment edges in DeFi across multiple blockchains. It provides a sophisticated platform to discover and construct high-potential, risk-adjusted portfolios. The system automates the journey from raw, multi-chain data (spanning 8 EVM chains via 1inch APIs) to actionable investment strategies. Data is aggregated, cleaned, and fed into an intelligent forecasting engine using quantitative models (Fourier analysis, GARCH for volatility) and technical analysis (RSI, MACD, Bollinger Bands). Assets are then intelligently scored and ranked based on collective signal strength. Optimized portfolios are constructed using Modern Portfolio Theory (MVO) with objectives like maximizing Sharpe Ratio, minimizing volatility, or maximizing return. The goal is to empower users to navigate DeFi markets with greater confidence and data-backed precision.

AlphaStream was built in 36 hours for ETHGlobal Prague.

Backend: Built with Python and FastAPI, it ingests token information and OHLCV price history from 1inch APIs across 8 EVM-compatible chains. This data is cleaned, standardized, and stored in MongoDB using Motor for async operations. The core forecasting logic in forecast/ uses Pandas, NumPy, Scikit-learn, and Statsmodels for quantitative models (Fourier, GARCH) and technical analysis (RSI, MACD). A custom Mean-Variance Optimization (MVO) module (mvo_portfolio.py) constructs portfolios. Services like one_inch_data_service.pyandmongo_service.py handle external integrations and database operations.

Frontend: Developed with Next.js 15 (React 19) and TypeScript, styled with Tailwind CSS and Radix UI components. Recharts is used for data visualization, and RainbowKit, Wagmi, and Viem manage wallet connections. Zustand handles state.

Integration: The Next.js frontend communicates with the FastAPI backend to fetch data and portfolio suggestions. MongoDB caching for OHLCV data and portfolio results ensures fast responses.

Partner Tech: The 1inch API was crucial for comprehensive multi-chain data aggregation and is planned for future trade execution via Fusion+. Hackathon Challenge: A key achievement was rapidly integrating these disparate systems—multi-chain data ingestion, complex financial modeling, and an intuitive frontend—within the 36-hour timeframe. The modular backend architecture with a clear pipeline (main_pipeline.py) and service-oriented design was essential for this rapid development.