I’m looking to build a fully autonomous trading algorithm for the US equities market. The core goal is to combine rigorous historical back-testing with real-time AI-driven decision making so the system can open, manage, and close positions without manual intervention. Key elements I need: • A Python-based trading engine that plugs into my existing .NET environment and leverages machine-learning libraries (TensorFlow, PyTorch, scikit-learn) for signal generation. • A robust back-testing module able to ingest years of tick and OHLCV data, run walk-forward validation, and produce detailed performance metrics (Sharpe, max drawdown, win-rate, exposure). • Live execution logic that routes orders to a US equities broker API, handles partial fills, slippage, and exchange-specific rules. • Risk controls: position sizing, portfolio-level VaR, hard stops, and circuit-breaker style kill switches. • Logging & monitoring dashboards (web-based) that show real-time P&L, open orders, and model health statistics. • Clear, well-commented code and concise documentation so I can audit, extend, and redeploy models quickly. I have data sources, hosting, and brokerage access ready; what I need is the algorithmic brain, the testing harness, and the production-grade glue to tie it all together. If you have a proven track record in Python algo-trading, machine learning, and .NET interoperability, I’d love to see how you would approach this build and refine it side by side with me through to a live launch.