Hft ea for mt4 mt5 live account

I have a High-Frequency Trading (HFT) Expert Advisor for both MT4 and MT5 that is primarily designed to trade US30 (Dow Jones Index). The EA performs consistently and profitably on demo accounts; however, when deployed on an IC Markets Raw or Standard live account, it begins generating losses despite what appear to be the same market conditions.

At this time, I do not have access to the source code (.mq4/.mq5). I can only provide the compiled EA files (.ex4/.ex5) for testing, evaluation, and behavioral analysis.

I am looking for a highly experienced MQL4/MQL5 developer with strong expertise in HFT execution, low-latency trading, slippage, spreads, order execution, broker execution models, and the differences between demo and live trading environments. Experience optimizing or developing US30 HFT or other latency-sensitive Expert Advisors is strongly preferred. If you know exactly the type of EA I'm referring to and already have a similar project with source code, or you've successfully solved this kind of live-vs-demo execution problem before, that would be a major advantage.

The goal is to determine why the EA behaves differently on IC Markets live accounts while trading US30. I need someone who can identify whether the losses are caused by factors such as execution latency, slippage, spread variations, requotes, order handling, broker-side execution, server limitations, trade timing, or other live-market conditions that are not present on demo accounts.

If possible, I would like the EA to be optimized, rebuilt, or redeveloped so its live-account performance closely matches its profitable demo performance.

Additionally, I would like an optional Enable/Disable Safety Mode that intelligently reduces unnecessary trading activity and limits excessive server requests (approximately 2,500 server messages per day) while maintaining the strategy's overall performance and execution quality.

If you have experience with HFT systems, latency arbitrage, ultra-low-latency execution, broker execution optimization, or similar US30 Expert Advisors, please describe your relevant experience and explain how you would approach diagnosing and solving this issue.

Because I do not have the original source code, I am also interested in discussing the most practical path forward, whether that involves reverse engineering the EA's behavior, rebuilding the strategy from scratch, developing a similar system based on observed logic, or using an existing HFT framework you already own and can adapt.