Pranav

Developed prepayment risk models on a proprietary residential mortgage portfolio of ~800,000 loan-level observations and ~€3B notional, working directly with the firm's internal dataset. Replaced a static constant parameter framework with a Markov-Switching model with time varying transition probabilities, capturing regime shifts in borrower behavior driven by macroeconomic factors. Implemented a two factor Hull-White interest rate model and Monte Carlo simulation framework to value embedded prepayment options at portfolio level, producing valuations 36% higher than the existing benchmark. Applied LASSO variable selection and 5 fold cross validation to identify the most predictive macroeconomic drivers from a high dimensional feature set.

Built Python-based valuation and simulation tools for 4 energy and technology sector companies, integrating live Bloomberg data to model financial performance across multiple volatility regimes. Conducted deep dive financial analysis covering earnings drivers, risk scenarios, and valuation multiples, delivering research outputs under tight deadlines in a fast paced institutional setting. Gained direct exposure to how tier 1 investment bank research teams translate quantitative analysis into actionable investment conclusions

Built predictive clustering and regression models on a proprietary dataset of 250+ firms for the Port of Rotterdam, identifying the structural drivers of innovation performance across a large industrial portfolio. Developed fully reproducible Python pipelines (pandas, scikit-learn, statsmodels) covering data cleaning, feature engineering, model training and evaluation. Conducted rigorous robustness checks including VIF, ANOVA, and out-of-sample validation to ensure model reliability. Translated complex analytical findings into clear, actionable recommendations for senior stakeholders