Abdelmalik — Machine Learning freelancer in Casablanca

🚀 Hello! I’m Abdelmalik, Quant Trader & Quantitative Researcher, building production-grade buy-side trading and pricing systems, from signal generation to live execution under real-market constraints. I design end-to-end quantitative engines that operate reliably with noisy or delayed data, microstructure effects, latency budgets, and strict forward-only decision rules. My work is desk-oriented: models are judged on robustness, execution quality, regime stability, and deployability. 📈 Quant Trading, Research & Derivatives Systematic Trading & Execution. Live strategies with microstructure-aware gating (spread in basis points, aggressor ratio, upticks). Regime-adaptive logic across trend, volatility and liquidity. Risk-based sizing using ATR, staged exits and VWAP-aware liquidation. Broker OCO orders, circuit breakers and idempotent execution to prevent double fills. Deterministic replays and ghost A\u002FB tests before production. 📊 Options & Exotic Pricing Pricing engines in C++ and Python using Black–Scholes, Heston, Dupire and Variance Gamma. Exotic structures including barriers and baskets. Numerical methods based on correlated and variance-reduced Monte Carlo, and PDEs using Crank–Nicolson with Rannacher smoothing. Volatility surfaces, smile reconstruction, robust proxy aggregation, walk-forward debiasing, affine scale correction and panel volatility guards. 📉 Validation & Results Strict walk-forward validation and multi-regime stress tests. R² between 0.83 and 0.92 on sector basket option pricers. RMSE below 1 basis point on barrier pricing grids. Sub-second C++ kernels with deterministic and replayable outputs. Stable pricing using 15-minute delayed options data. 🧠 Machine Learning (Production) Models built with Sklearn, XGBoost, PyTorch and TensorFlow. Probabilistic calibration using Platt scaling and isotonic regression, with Brier score and ECE monitoring. Drift detection using PSI, class-imbalance handling and robust cross-validation. GPU acceleration with CUDA and CuPy. Deployment via FastAPI and Streamlit. ⚙️ Quant Infrastructure High-performance C++17 cores using OpenMP, AVX and PyBind11, with parallel Monte Carlo and Cholesky correlation. Market data pipelines using Polygon with minute-level options and level 2 data. Production observability with Prometheus and Grafana SLOs, Redis-based event architecture, atomic snapshots and audit-safe logs. 🎓 Training: Columbia, Yale, Stanford, Wharton, NYIF 🌍 Availability: Remote worldwide or on-site Casablanca 🤝 Open to buy-side quant trading, systematic execution and advanced pricing or risk teams