I want to commission an AI-driven trading bot that focuses on pure scalping logic. The core job is to design, train and deploy a system that can enter and exit positions within minutes, extracting small, repeated gains throughout the trading day. Market scope The bot should specialise in US30, Gold and NAS100. While these aren’t classic “forex pairs”, I actively trade them through my broker’s currency-denominated contracts, so treat them exactly as you would any high-liquidity FX instrument. Decision engine All entries and exits must be generated from technical indicators only—no news feeds or sentiment data. You are free to combine indicators (e.g., EMA cross, RSI, Bollinger Band breakout, VWAP, order-book metrics, etc.) so long as the final logic stays true to fast-paced scalping principles. Key expectations • Back-test the strategy on tick or one-minute data to prove edge over at least two years. • Provide an optimisation notebook or script so I can fine-tune parameters later. • Package the bot for live execution on MetaTrader, cTrader or a Python-based API—whichever best preserves low-latency execution; recommend what suits a scalper. • Include a safeguard module for max drawdown, daily stop, news blackout periods and connection loss. • Deliver full source code, a brief setup guide and sample trade logs that match the back-test KPIs. Acceptance criteria 1. Net profit and Sharpe ratio outperform simple buy-and-hold on all three instruments over the test window. 2. Maximum intraday drawdown never exceeds 3 %. 3. Average trade duration remains under 15 minutes. If you have prior experience writing high-frequency or scalping bots and can demonstrate real-world slippage management, let’s talk.