Shardul Mehetar

Developed systematic equity and ETF signals using Python and statistical analysis across momentum, mean-reversion, and market-regime hypotheses. Built backtesting workflows including transaction costs, slippage, volatility targeting, and portfolio constraints, and evaluated strategy behavior across multiple historical regimes. Validated alpha signals using out-of-sample, walk-forward, and forward testing while applying lookahead-bias detection and signal-decay analysis for robustness. Productionized research workflows with Git and Linux/UNIX and collaborated with quant developers and portfolio managers to monitor daily PnL, exposures, and strategy behavior.

Developed quantitative equity signals in Python using Pandas/NumPy and scikit-learn, testing cross-sectional factor hypotheses across large-cap universes and sector groups. Built SQL and Bloomberg-based screening workflows to analyze factor exposures, performance attribution, and portfolio risk metrics. Researched momentum and factor relationships and applied feature engineering plus statistical validation to identify persistent signals. Automated recurring research and reporting workflows with Python/SQL/Git, reducing manual processing and improving reproducibility; validated signals with out-of-sample and robustness testing before presenting findings to portfolio stakeholders.

Developed retail PD models using Logistic Regression and XGBoost (AUC-ROC 0.87) on imbalanced credit data, reducing misclassification-driven capital error by 16% using asymmetric loss calibration. Engineered credit-risk features from transactional and bureau data using Python/Pandas/SQL with missing-value handling, outlier analysis, and transformations to improve stability. Designed a Champion–Challenger framework using PSI/CSI to trigger justified retraining, compressing review cycles from weekly to daily with zero manual intervention. Validated performance with cross-validation and out-of-time testing, using KS, AUC-ROC, and population stability analysis; automated development and validation workflows with Python/SQL/Git for standardized, reproducible reporting.