System Architecture
Algo Master: Algorithmic Trading Platform Overview
A comprehensive 3-tier microservices architecture powering intelligent algorithmic trading
6 Services 45+ API Entities 6 Brokers 100+ ML Features 3 Domains
Table of Contents
01 Platform Architecture 04 Technology Stack 02 Three-Tier Design 05 Data Flow 03 Service Communication 06 Key Modules
Section 01
Platform Architecture
The system is organized as a three-tier microservices architecture with dedicated services for data, compute, and presentation.
System Architecture Diagram
React Frontend Port 3030
Trading UI Analytics Dashboard Real-time Charts Options Scanner Strategy Builder
▼
Strapi Middleware Port 1343
API Gateway Authentication CORS Handling Request Proxy Cron Tasks
▼
Python FastAPI Port 8100
Order Execution Greeks Engine ML Models Strategy Logic Backtesting
WebSocket Server
Port 8766
PostgreSQL
Port 5432
Redis
Port 6379
Section 02
Three-Tier Design
Each tier is independently deployable with clear responsibilities and well-defined interfaces.
R
React Frontend
The presentation layer providing a rich trading experience with real-time data visualization and interactive analytics.
MUI v6 TradingView Charts Redux Socket.io Capacitor Mobile
S
Strapi Gateway
The middleware layer handling authentication, request validation, response formatting, and proxying to the Python backend.
Request Validation Response Formatting Auth Middleware CORS Cron Tasks
P
Python Backend
The compute layer powering all trading logic, quantitative analysis, machine learning, and broker integrations.
FastAPI Async Multi-Broker Greeks Engine ML Pipeline WebSocket Streaming
Section 03
Service Communication
Request lifecycle from user interaction to response rendering, flowing through all three tiers.
User Action
→
React App
→
Strapi Proxy
→
Python API
→
Broker / DB
→
Response
→
UI Update
Section 04
Technology Stack
A modern stack chosen for performance, developer experience, and production reliability.
| Layer | Technology | Purpose |
|---|---|---|
| Frontend | React 18 |
Component-based UI with concurrent rendering |
MUI v6 |
Material design component library | |
TradingView |
Professional charting with custom indicators | |
Redux Toolkit |
Global state management | |
Socket.io |
Real-time data streaming to UI | |
| Middleware | Strapi v4 |
Headless CMS and API gateway |
Bookshelf ORM |
Database abstraction layer | |
Axios |
HTTP client for Python API proxy | |
| Backend | FastAPI |
Async Python web framework |
SQLAlchemy |
ORM and database toolkit | |
Pandas / NumPy |
Data manipulation and numerical computing | |
Scikit-learn / XGBoost / PyTorch |
Machine learning and deep learning models | |
Redis / aioredis |
Async caching and pub/sub messaging | |
| Database | PostgreSQL |
Primary relational database |
| Cache | Redis |
In-memory cache, session store, pub/sub |
| Infra | PM2 |
Process manager for Node.js services |
Docker |
Containerization and deployment | |
Uvicorn |
ASGI server for FastAPI |
Section 05
Data Flow
End-to-end data pipeline from market data ingestion through ML prediction to order execution.
Market Data Ingestion
Broker feeds stream via WebSocket, normalized and stored in the MarketData table with real-time tick updates for LTP, volume, and open interest.
Feature Engineering
Over 100 features computed from OHLC data, open interest changes, Greeks surfaces, technical indicators, and time-of-day signals.
ML Prediction
12 signal sources including LSTM, Transformer, and ensemble models aggregated into a unified trading score ranging from 0 to 100.
Strategy Execution
Orders placed through a multi-broker factory pattern supporting Alice Blue, Upstox, and four additional brokers with unified order interfaces.
Monitoring & Analytics
Real-time P&L tracking, Greeks exposure dashboard, dealer flow analysis, and strategy performance metrics rendered on the frontend.
Section 06
Key Modules
Core functional domains of the platform, each covered in dedicated deep-dive blogs.
Δ
Options & Greeks
Black-Scholes pricing engine with real-time IV surface computation, dealer flow analysis tracking institutional positioning, and strike-level Greeks visualization.
Black-Scholes IV Surface Dealer Flow OI Analysis
⚙
ML & Deep Learning
LSTM and Transformer models for time-series forecasting, XGBoost signal aggregation across 100+ features, and unified scoring pipeline.
LSTM Transformer Signal Aggregation 100+ Features
⚙
Trading Strategies
Automated straddle management, crush recovery systems, martingale position sizing, and momentum-based option scanning with configurable parameters.
Straddle Crush Recovery Martingale Momentum Scanner
⏲
Backtesting
Historical straddle backtester with multi-leg support, derivatives simulator with intraday time-of-day modeling and IV surface prediction.
Straddle Backtester Derivatives Simulator IV Prediction Historical Replay
Reference
Port Mapping
Service endpoints and their network protocols for local development and production deployment.
| Port | Service | Protocol |
|---|---|---|
3030 |
React Frontend | HTTP |
1343 |
Strapi Middleware | HTTP |
8100 |
Python FastAPI | HTTPS (self-signed) |
8766 |
WebSocket Server | WSS |
5432 |
PostgreSQL | TCP |
6379 |
Redis | TCP |
Algo Master Platform · Algorithmic Trading System
Top comments (0)