Rayan Chatterjee

Background

About Rayan Chatterjee

I'm a Stanford-trained Quantitative Researcher with a PhD in Physics, specializing in mathematical modeling, statistics, and high-performance simulation. My expertise spans model development, model validation and sensitivity analysis —skills central to credit and market risk modeling. I’ve built robust machine learning pipelines for credit default prediction using real-world financial data—handling missing values, resolving multicollinearity with VIF, training logistic regression models, correcting class imbalance via SMOTE, and interpreting results with SHAP. This helped uncover key default predictors, eliminate spurious predictions, and improve model generalization. I've also conducted in-depth VaR and stress testing on a 4-asset portfolio (AAPL, MSFT, SPY, TLT), revealing how extreme events like COVID spikes affects risk estimates. I also stress-tested credit risk models, showing how small feature shifts can trigger disproportionate risk spikes—highlighting real-world model fragility and sensitivity. Furthermore, I have also explored how risk evolves over time using Static, EWMA, and GARCH(1,1) VaR models. Backtests with Kupiec’s POF and Christoffersen’s tests confirmed GARCH as the most accurate in capturing volatility clustering—demonstrating how dynamic models lead to more resilient risk estimates. During my postdoctoral research at Stanford and in the University of Pittsburgh, I have developed and validated mathematical models for complex dynamical systems. For model development, I fitted time-series data using p-values, adjusted R², AIC, and performed sensitivity analysis of the model response to parameters. Earlier, during my PhD at TIFR, I developed parallel solvers for nonlinear PDEs and analyzed time-series data using skewness, kurtosis, and correlation functions. During my masters at IIT Kanpur, I applied PCA (aka POD) to turbulent flows. I bring a research-driven, implementation-focused mindset to quantitative risk—combining statistical rigor, coding fluency, and practical insight to solve complex risk problems.