Forex Statistical Arbitrage Algorithm Development

I want to build a statistical-arbitrage engine that trades the major forex pairs—think EUR/USD, GBP/USD and the rest of the high-liquidity set. The core objective is to exploit short-term mean-reversion and cointegration opportunities, all fully automated from signal generation through execution.

Here is what I need from you:

• Strategy logic coded in a language suited for low-latency connections to my broker’s API (Python with NumPy/Pandas is fine, but I’m open to C++ or a mixed approach if latency becomes critical).
• Robust data-handling: live tick or one-second data ingestion, plus a pipeline for historical price pulls so we can back-test properly.
• Back-testing framework that reports Sharpe, max drawdown, win rate and trade distribution, with parameter optimisation built in.
• Risk controls baked into the code—position sizing, dynamic stop-loss, and circuit breakers that pause the strategy if drawdown limits are breached.
• Deployment script so I can run the bot 24/5 on a VPS with automatic log rotation and email/SMS alerts for key events.

Acceptance criteria
– Strategy replicates at least 95 % of back-test logic in live forward-testing on a demo account over two trading days.
– Slippage model included in back-tests and no unexplained live/expected P&L drift larger than 0.3 %.
– Clean, well-commented source code delivered via Git repository together with a concise README.

If you have prior experience building profitable stat-arb models in the forex space, especially on major pairs, I’d love to see examples of your work or performance metrics.