Back to Home
Project 02

Trading Bot

Algorithmic execution · Realtime signals

An algorithmic trading engine with backtesting, risk management, and live execution. Composable strategies, multi-timeframe analysis, and a live monitoring dashboard.

Role
Developer & strategist
Status
Local-only build
Period
2023 — Now
Domain
Fintech · Markets
PythonpandasccxtTimescaleDBPlotlyAsync WebSockets
The Problem

Most trading bots either over-fit a single strategy or lock the operator into a black box. The goal was something composable, observable, and safe enough to run live.

Approach

Treat the bot as an engine with three concerns — data, strategy, execution — each independently testable, with a backtest lab that mirrors the live runtime exactly.

Architecture

Data plane

ccxt-driven feeds normalized into TimescaleDB; multi-timeframe pandas resampling for indicators.

Strategy layer

Composable strategy modules with a thin contract: signals in, sized orders out. A risk engine wraps every signal.

Execution

Async order router with idempotent keys; sub-40ms execution to market.

Observability

Plotly dashboard streaming indicator state, PnL, and the risk envelope live.

Outcomes
  • 12 composable strategies covering trend, mean-reversion and event-driven setups.
  • Sub-40ms execution latency with a 24/7 uptime design.
  • Backtests reproduce the live runtime — no surprise drift.
Stack
PythonpandasccxtTimescaleDBWebSocketsPlotlyAsync I/O
Screenshots
Monitoring dashboard — live indicator state, PnL, and risk envelope.