I have an MVP that copies trades from MT4 to Bitget USDT-M Futures. It works, but it’s still a proof-of-concept. I need an experienced dev to harden it for live trading. Top priority Rewrite the close logic so Bitget positions close with 100% accuracy (handle partial fills on exchange side, race conditions, and rapid markets). Preserve hedging (simultaneous long & short) and use reduce-only for all closes. Fixed-USDT sizing per trade (e.g., 1/5/10 USDT). No partial closes by design. Additional polishing Tighten symbol mapping (MT4 symbol → correct Bitget contract) via config. Robust error handling & retries (timeouts, network drops, 429 throttling, signature skew). Move key settings (API keys, mapping, fixedUSDT) to a clean config file (or simple UI). Optional performance pass (session reuse, pooled HTTP, clock sync). Acceptance criteria 1. MT4 BUY → Bitget open_long (market); MT4 close → reduce_only long (full). 2. MT4 SELL → Bitget open_short (market); MT4 close → reduce_only short (full). 3. Parallel BUY & SELL on the same symbol open and close independently. 4. No duplicate opens/closes under retry. 5. Latency target: p50 ≤ 300 ms, p95 ≤ 600 ms on an EU VPS (FFM/AMS). Deliverables Refactored, well-commented source (MT4 EA + bridge in Python/Node). Short test script / walkthrough proving accurate open/close on ≥2 symbols. Changelog of fixes and improvements. Quick-start guide to redeploy on my VPS in minutes. Stack & skills MT4/MQL4, Bitget REST & WebSocket APIs, rate-limit savvy. Security: trade-only API key, no withdraw; IP whitelist preferred.