Yanis B.

As a French Quantitative Algorithmic Trader with dual MSc degrees (Quantitative Finance; Mathematics Applied to ML/DL) and 4+ years of experience, I specialize in building and optimizing end-to-end algorithmic trading systems with a strong focus on backtest quality, strategy design, and deployment readiness. I excel in translating advanced Quantitative Research and Trading Strategies into production-grade Python trading pipelines that deliver consistent, risk-adjusted returns across different asset classes. Core Expertise: Robust Backtesting & Validation - Designed rigorous k-fold walk-forward validation protocols to ensure model generalization and robustness. - Specialized in high-fidelity backtesting frameworks (5+ years of tick/minute data) with transaction costs, slippage, and execution modeling. - Optimised take profits, stop losses and size to improve returns, Sharpe Ratio and Calmar Ratio. Algorithmic Trading Systems in Python : - Proficient in integrating ML/DL models, RL agents, and signal ensembles into live-ready trading codebases. Strategy Design & Innovation - Experienced in alpha factor research, systematic strategy design, and multi-horizon forecasting. - Skilled at combining quantitative finance models (Heston, SABR, volatility surfaces) with modern ML/DL architectures (Transformers, GNNs, RL). Machine Learning & Reinforcement Learning for Trading : -Built RL-based trading agents (PPO, DQ-Net + TFT) achieving Sharpe 1.95 and 20% alpha on BTC spot forecasting. - Developed multi-model ensembles (TFT + Graph Neural Networks) for volatility and alpha prediction. Data Engineering & Feature Integration - Engineered complex feature sets from OHLCV, on-chain metrics, sentiment data (Twitter, Reddit, News), and cross-asset signals. Key Project Highlights: -BTC Perpetual Futures – Live Algo: Designed and deployed a Python strategy achieving Sharpe 1.6 and Calmar 1.6 after costs, with 15% outperformance vs benchmark returns over live trading horizon. - SPY Futures – Multi-TF FVG Strategy: Built a multi-timeframe Fair Value Gap model with positive Sharpe/Calmar, demonstrating robustness and adaptability to equity index futures. -Developed RL agents ( DQ-Net with TFT) with Sharpe 1.95 Calmar 1.6 and 20% alpha in crypto spot forecasting. Technical Toolkit: Languages: Python ML Frameworks: PyTorch, TensorFlow, scikit-learn, Optuna, Vectorbtpro, RLTrading Data Tools: pandas, numpy Domains: Equity index, Crypto/Altcoin Markets, Time-Series Forecasting, RL, Risk Management Education: - MSc Quantitative Finance (Paris Saclay) - MSc Mathematics Applied to Machine Learning & Deep Learning (ENSIIE)