AI-Powered Multi-Asset Trading Bot (Crypto + XAU/USD) — Python, Hetzner Deployment

Here's the full posting in English: Title AI-Powered Multi-Asset Trading Bot (Crypto + XAU/USD) — Python, Hetzner Deployment Project Description I'm looking for an experienced developer to build an automated trading bot that runs 24/7 on my Hetzner server and makes trade decisions using an AI-driven decision layer. The bot must support both crypto exchanges and XAU/USD (gold). For XAU/USD I already have a clear idea of the strategies I want pre-built into the system (details below), while crypto can use a more general, configurable strategy framework. This is not a "magic profit bot" request — I understand markets. I need clean, well-architected, production-grade software that I can extend, monitor, and trust to run unattended. What the Bot Should Do (Scope) Connect to: Crypto: [Binance / Bybit / OKX — pick one] via CCXT (or native API) XAU/USD: MetaTrader 5 (MT5) via the Python MT5 library / a broker bridge AI decision layer: the bot evaluates market data and signals, then an AI component produces a trade decision (enter / exit / hold / size). I'm open to the approach — LLM-based reasoning over structured signals, an ML classifier, or a hybrid — and I expect you to recommend what's realistic and testable. Pre-built XAU/USD strategies baked in as selectable modules, including: EMA crossover entries Support/resistance level detection Pin bar / price-action confirmation (Each strategy toggleable and parameterizable via config) Risk management: position sizing, max daily loss, stop-loss / take-profit, max concurrent positions. Backtesting on historical data before going live, plus a paper-trading / dry-run mode. Logging, alerting, and a basic status dashboard (Telegram alerts and/or a simple web panel are fine). Technical Requirements Language: Python (preferred). Deployment: Dockerized, deployed on my Hetzner server. Must be reproducible (docker compose up), with clear env-based configuration. Architecture: modular — adding a new strategy or a new exchange should not require rewriting the core. Reliability: auto-restart on failure, persistent state, safe handling of API/connection drops. Secrets: API keys handled securely (env / secrets, never hardcoded). Documentation: README covering setup, configuration, adding strategies, and running backtests. Skills Required Python, CCXT, MetaTrader 5 (MT5), algorithmic / quant trading, Docker, REST/WebSocket APIs, backtesting frameworks, and experience integrating AI/ML into trading or decision systems. Deliverables Working bot deployed and running on my Hetzner server. Source code in a private Git repo. Backtest results for the XAU/USD strategies. Documentation as above. A short handover call walking me through the architecture. To Apply, Please Include A short example of a similar trading bot or automation system you've built (CCXT or MT5 a strong plus). Your suggested approach for the AI decision layer. Estimated timeline and your rate.