Introduction
The backend of TradeSim is developed using:
- Python – The core language for backend logic and data processing.
- FastAPI – A high-performance asynchronous web framework used to expose REST and WebSocket APIs.
- scikit-learn – Used for training and deploying the machine learning model that predicts midpoint prices.
Key Responsibilities
-
🔌 API Services
Serves REST and WebSocket endpoints for trade simulation, orderbook updates, and prediction metrics. -
🧠 Machine Learning Model
A trained scikit-learn model predicts the midpoint price 60 seconds into the future using real-time orderbook data. -
⚙️ Trade Simulation Engine
Processes incoming market order requests, simulates fills based on current orderbook state, and returns detailed trade execution metrics. -
🔄 Real-time Data Handling
Ingests live market data (e.g., from OKX), updates the orderbook cache, and pushes real-time updates to connected clients via WebSockets.
Technologies
| Component | Tech Used | Description |
|---|---|---|
| Web Framework | FastAPI | Async framework for HTTP and WebSocket APIs |
| Machine Learning | scikit-learn | Model training and inference for predictions |
| WebSockets | FastAPI WebSocket | Real-time data push to frontend |
| Market Data Provider | OKX API (external) | Source of live orderbook snapshots |
The backend is designed to be modular, scalable, and ready to integrate more exchanges, assets, and advanced ML capabilities in future updates.