Youness E.

I help traders and funds find out whether a strategy really works, then run it: rigorous backtests, ML trading signals, and live performance dashboards. Kaggle Grandmaster (Datasets & Notebooks) · Top Rated · 100% Job Success · 7+ years in financial ML WHAT I DO - Strategy backtesting & validation: QuantConnect, vectorbt, backtesting.py and in-house walk-forward frameworks. Realistic fees and slippage, out-of-sample testing, leakage and overfitting checks, so you see real performance, not a curve-fit. - Trading dashboards: Bloomberg-style dashboards tracking equity curves, drawdowns, Sharpe/Sortino/Calmar, monthly returns and live signals, deployed on your server with access control. - Alpha research & ML signals: factor research, regime detection and forecasting for futures, crypto and equities (1-min to daily). XGBoost/LightGBM, and deep learning only when it beats simpler models. - Market data pipelines: automated collection, cleaning and storage of equity, crypto, on-chain and news data, including survivorship-bias-free universes. RECENT WORK - Multi-strategy BTC suite: six systematic strategies backtested on QuantConnect, with an access-controlled live dashboard whose orders, trades and equity reconcile with QuantConnect's accounting. - Bloomberg-style quant terminal: a live dashboard covering five strategies (equity curves, drawdowns, risk metrics, monthly heatmaps, live signal page), running 24/7 on a VPS with automated daily data refresh. - Institutional equity data pipelines: Russell 1000 historical universe including delisted stocks via the FactSet APIs, plus FactSet and StreetAccount news headline pipelines. - Leakage audit: reviewed a portfolio weighting function for look-ahead bias and documented the upstream risks. - On-chain BTC research: tested exchange, custody and stablecoin flow features with walk-forward models and regime analysis, with an honest verdict on what is and isn't tradable. HOW I WORK Fixed-scope milestones, clean reproducible code, and a plain-English report explaining what the results mean and where the risks are. Live dashboard demos available on request. STACK Python (pandas, NumPy, scikit-learn, statsmodels, PyTorch), QuantConnect, vectorbt, Optuna, FactSet APIs, SQL, Streamlit, FastAPI, Scrapy/Selenium, Linux/VPS deployment. Send me your strategy idea or dataset, and I'll reply within 24 hours with how I'd test it.