Lead Quant Researcher – Production ML Pipeline to Reverse-Engineer Options Trading Logic

Project OverviewWe possess 5 years of historical options trading signal data (comprising Index and Stock options). The core objective is to reverse-engineer these signals to uncover their exact underlying algorithmic, technical, or structural trading logic. We require a top-tier Quantitative Researcher and Systems Architect to build a production-grade Python framework. This system must do two things:Reverse-Engineer Logic: Ingest historical raw signal data, map it to historical market data, and output the precise rules, chart patterns, and mathematical conditions that triggered those signals.Production Inference & Feedback Engine: Run as a standalone production tool. When a completely new dataset of raw market data is fed in, it must generate inference on likely signal triggers. Conversely, when new signal data is fed in with observed outcomes, the system must ingest it as a feedback loop to dynamically optimize and update its logical models. Technical Scope of Work1. Production Code Architecture (Python)Dual-Mode Engine: Build a modular, production-ready system capable of both bulk historical training and real-time/batch inference.Closed-Loop Feedback System: Implement a pipeline where new market data and subsequent trade outcomes are ingested to continuously score, refine, and adapt the discovered logic models.2. Deep Price Action & Structural Pattern RecognitionMulti-Timeframe Chart Analysis: Scan and identify candle structures and macro chart patterns (e.g., Head & Shoulders, Cup & Handle, Double Tops/Bottoms, Flags) across all timeframes (1m to Daily).Institutional Order Flow: Track "big money" footprints, detecting institutional liquidity pools, block deals, and volume surges immediately preceding the signals.Market Friction Anomalies: Account for session gap-ups/gap-downs, weekend risk mitigation, and the structural impact of long market holidays on accelerated option premium decay.3. Derivative Dynamics & Market Maker BehaviorStop-Loss Hunting Mechanics: Investigate if recommended stop-losses were mathematically placed to exploit retail liquidity zones or key Open Interest (OI) walls frequently targeted by market makers.True Profitability & Friction Audit: Run a baseline verification of signal success rates against actual historical tick/1-minute data, explicitly calculating slippage based on order book depth at the exact millisecond of signal generation.Risk-Reward Asymmetry: Reverse-engineer how entry, target, and stop-loss levels were derived (e.g., dynamic ATR tracking, fixed percentage, or volatility-adjusted bands).4. Event-Driven & Macro AnomaliesRegime Classification: Isolate signal behavior during high-impact windows (Union Budget, RBI Policy days, corporate earnings, global market shocks).Expiry & VIX Interaction: Map signal decay dynamics relative to India VIX levels and proximity to weekly/monthly contract expiries.5. Advanced Mathematical Modeling & Machine LearningPattern Recognition via Computer Vision: Implement Convolutional Neural Networks (CNNs) on transformed financial time-series graphs to check if visual chart patterns triggered the signals.Predictive Modeling: Utilize Deep Neural Networks (DNNs) and ensemble ML algorithms (XGBoost, Random Forests) to classify features that yield high-probability signals.Risk Stress-Testing: Run Monte Carlo simulations to evaluate strategy drawdown distributions, risk-of-ruin, and optimal position sizing (e.g., Kelly Criterion). Required Technical Stack & QualificationsExperience: Minimum 3–5 years as a Quantitative Researcher or Algorithmic Systems Architect with explicit domain expertise in Indian Derivatives (NSE).Languages/Frameworks: Expert-level Python (Pandas, NumPy, Scikit-Learn, PyTorch/TensorFlow, TA-Lib, Backtrader/Zipline).Data Mastery: Demonstrated experience handling massive NSE data structures, including historical option chains and order book data.Architecture: Proven experience building production-grade ML pipelines with clean code principles, robust error logging, and configuration management. DeliverablesThe Signal Audit Report: Detailed data analysis breaking down true historical profitability, slippage impacts, and risk-reward profiles.The Logic Blueprint: Document defining the exact combination of price action, macro regimes, and option greeks that trigger the signals.Production Python Codebase: A fully functional, modular pipeline containing:Data pipelines for processing NSE market data and options chains.The CNN and ML feature-extraction modules.The Inference Engine for running on new data.The Feedback Loop module for continuously updating models with new outcomes.