Predicted 2024 NYC Taxi Demand by deploying Seasonal ARIMA, Exponential Smoothing, and Prophet models with MAPE of 2.14%, optimized driver schedules and rider wait times, translating to 65% increase in operational efficiency

Engineered a Credit Card Fraud Detection logic by training ensemble models (RF, CatBoost, XGBoost, LightGBM) with a test accuracy of 78.34% at 3% FDR. Identified over 70% of fraudulent applications, translating to $12.2M in annual savings

Developed a facial emotion recognition system by training a CNN model on 54K images and used Kernel SVMs for classification, achieving sub-5% loss throughout training. Also, visualized performance with F1-score vs. epochs

Proficiently managed an NLP pipeline by Implementing tokenization, stemming, and diverse feature extraction techniques such as Bag Of Words, TF-IDF, and Word2Vec, attaining an accuracy of around 90%. Deployed Logistic Regression and Support Vector Machine (SVM) to classify the text

Utilized Tableau for Netflix growth and revenue analysis, identifying key drivers including subscriber growth, content consumption patterns, and geographic trends. Analyzed factors such as subscription plans and content acquisitions to optimize strategies, resulting in enhanced revenue streams

Developed a K-Means clustering model, employing the Elbow Method optimization technique, and performed Customer Demographic Analysis by extracting their order purchase details. Leveraged PCA analysis to deepen insights and refine understanding of customer segmentation dynamics.