My Portfolio

Empirical study comparing PPO and TRPO across MuJoCo environments using 50+ seeds and multi-metric evaluation: IQM, AUC, MAD, CVaR, and hyperparameter sensitivity.

Unsupervised learning on Google attraction reviews. Used PCA and k-means clustering to uncover visitor groups and behavioral patterns.

Predicting 2025 protest activity using bootstrapping and Monte Carlo simulation on historical datasets.

Machine learning classification of dry bean species using SVM (RBF kernel). Achieved 91% accuracy and strong feature-level insights.

Epidemiological modeling of syphilis in Alberta under various public-health intervention scenarios.

Message me if you'd like the full code base.

Time series modeling using ARIMA, SARIMA, and smoothing methods to forecast quality behavior over time.