Shubham Godse

Versatile and results-driven Quant Developer with a strong foundation in financial engineering, statistical modeling, and machine learning applied to capital markets. Skilled in designing and deploying systematic trading strategies, predictive risk models, and real-time analytics pipelines across equities, FX, and credit domains. Demonstrated success in backtesting alpha-generating signals, optimizing Sharpe ratios, and automating data workflows for institutional-grade scalability. Proficient in Python, quantitative research, and model validation techniques aligned with front-office trading teams. Passionate about bridging financial theory with robust, production-level code to deliver actionable insights and resilient trading logic in fast-moving markets.

• Bankruptcy Risk Intelligence EngineDeveloped an XGBoost-based classification model leveraging financial ratios (ROE, CR, D/E) with calibrated thresholds to assess bankruptcy risk. Achieved 95% ROC-AUC and deployed in a live credit approval pipeline to support automated lending decisions.
• ML-Driven FX Strategy SimulatorEngineered a signal classification system for EUR/USD trading with dynamic position sizing and regime-aware adjustments. Walk-forward validation confirmed +21% annualized returns with controlled drawdown, demonstrating real-world trade viability.
• LLM-Based CAM Report GeneratorBuilt a structured NLP pipeline to extract key financial fields and generate credit commentary using Mistral, including token usage optimization and custom prompt tuning. Led to a 60% reduction in analyst turnaround time for CAM report generation.
• Quant Sentiment Heatmap EngineDesigned a hybrid relevance and sentiment model combining fuzzy matching, semantic filtering, and FinBERT sentiment scoring to rank sectoral news by bullishness. Used in an early risk alerting system for credit and investment teams.

• Prompt Engineering, Coursera, 01/01/24
• NPTEL Machine Learning, IIT Madras, 01/01/24
• IBM Data Analysis, Coursera, 01/01/23

• Achieved a 28% increase in Sharpe ratio by designing regime-aware signal calibration techniques, enhancing the robustness and profitability of ML-driven trading strategies.
• Delivered a production-ready LLM-based financial analytics solution that was successfully deployed by credit teams to automate balance sheet interpretation and portfolio commentary.
• Led the development of low-latency deployment infrastructure for quant ML models, reducing model response time by 30% and improving scalability for real-time execution environments.