Back to Works
AI Quantitative Trading

Smart AI Trading System

Solo-built automated crypto trading system with multi-service backends (exchange streams, on-chain monitors, macro/ML, AI engine) plus a trading engine with position/state sync and recovery. Includes WebSocket-based monitoring UI and log streaming.

Smart AI Trading System

Tech Stack

PythonFastAPIReactTypeScriptPostgreSQLDockerWebSocket

Year

2024 - Present

Executive Summary

The Challenge

Had to keep ingestion→analysis→execution pipelines reliable under continuous streaming data, while ensuring position consistency and recovery after restarts. Also needed operational visibility (service logs, status, event timelines) to debug long-running operation.

The Solution

Separated services by responsibility and connected them via trigger monitoring and a trading engine. Implemented startup recovery that syncs exchange positions and restores active sessions, plus log capture (file + in-memory) and WebSocket-driven dashboards for live monitoring.

Technical Documentation

Methodology & Implementation

Overview

  • Solo-built AI-powered crypto trading system combining real-time ingestion, multi-agent analysis, and automated execution.

Key Components

  • Robust ingestion/storage pipelines across heterogeneous data sources
  • Trigger-driven multi-agent analysis orchestration
  • Trading engine with position/state sync + restart recovery
  • Log capture + WebSocket monitoring for operational observability