Back to projects
Systems / FintechResearch

Trading Operating System

Modular algorithmic trading infrastructure with ICT/SMC analysis and prop-firm risk rules

TypeScriptNext.jsC#PythonTimescaleDBMT5
View source

Problem

Most retail algo-trading setups hardcode a single strategy directly against a broker API, with no separation between signal generation, risk validation, and execution — which makes the system impossible to test, extend, or reason about safely.

Solution

A modular cockpit and backend: a TypeScript ICT/SMC engine generates signals, a dedicated risk engine evaluates them against prop-firm-style rules before anything is allowed through, and a C# gateway translates approved decisions into an MT5 wire protocol — each stage documented with its own architecture decision record and covered by tests.

Architecture

Market data (MT5 observer)
ICT/SMC strategy engine
Risk engine (prop-firm rules)
Execution bridge → MT5 gateway
TimescaleDB persistence + cockpit UI

Key features

  • ICT/SMC signal-generation engine, documented in its own ADR
  • Standalone risk engine enforcing prop-firm-style drawdown and exposure rules
  • Execution bridge translating approved signals into MT5 commands via a C# gateway
  • TimescaleDB persistence for market data and audit trail
  • Cockpit UI: signals, positions, risk, backtests, replay, journal — one view per concern