Chunshen L.

𝐀𝐛𝐨𝐮𝐭 𝐌𝐞 I build algorithmic trading systems that are validated, production-ready, and actually trade. My own model runs live across 28 large-cap equity tickers: 80.7% win rate, Sharpe 4+, max drawdown 1.27%, validated over 89 days of live paper trading on Alpaca. Not a backtest. A live system I built and run myself. If you're working with someone who can say that with receipts, you're in the right place. 𝐖𝐡𝐚𝐭 𝐈 𝐁𝐮𝐢𝐥𝐝 Signal Models: Neural-network long/short classifiers trained on intraday price, volume, and macro features. Walk-forward validated, statistically tested, not curve-fit. Execution Infrastructure: Bracket orders, ATR-based stops, dynamic Kelly position sizing, full broker API integration (Alpaca). Regime Filtering: VIX/SPY macro overlays that scale or pause exposure based on market conditions. One model behavior for CALM markets, a different one for CRISIS. Backtesting & Validation: Out-of-sample testing, Sharpe attribution, p-value reporting. I'll tell you honestly whether your edge is real. System Audits: If your existing algo is underperforming, misfiring, or just feels off, I can diagnose it and tell you exactly why. 𝐒𝐭𝐚𝐜𝐤: Python, PyTorch, scikit-learn, XGBoost, deep learning (time-series), Alpaca API, pandas, NumPy.