Machine Learning Intern
• Developed and implemented various machine learning models, including a custom bagging neighbor model, achieving a 77% validation balanced accuracy for a cardiac catheterization project.• Utilized SMOTE and SMOTE-ENN techniques to address class imbalance, improving model performance by 15%.• Applied dimensionality reduction techniques (PCA, Isomap) and optimized model performance through cross-validation and Bayesian optimization.• Developed a PyTorch neural network model to predict adverse heart conditions, achieving 85% accuracy through advanced optimization techniques and RayTune for hyperparameter tuning.