Used national health and nutrition examination survey Data of 6329 adult subjects to implement ML classifiers that determine if a subject is physically active or not: \- Analyzed features that improve physical activity level based on gender and age groups using SHAP (SHapley Additive exPlanations). \- Re-sampled unbalanced dataset using different algorithms from imbalanced-learn API. \- Performed Hyper-parameters tuning using Sklearn GridSearch API to improve model metrics. \- Created women and non gender specific physical activity classifiers. SVM, XGboost and Random Forest . Obtained AUC metric of 0.81. Published a paper "Who is physically active? :Classification and analysis of Physical activity using NHANES data" in the IEEE SSCI conference

\- Improved virtual machine VM issue detection and resolution time by 90% by implementing a Random Forest classifier that classifies servers by current service/dept user. \- Created a webpage where users could interact with the classifier \- Optimized server configuration management process following ITIL - V3 best practices by implementing a new VM problem resolution process

\- Designed (UML) and implemented a recruitment website using PHP, HTML, CSS and SQL allowing candidates to fill their profile and upload their resume. \- Designed and developed a database to keep track of scientists and their expenses when traveling and attending different conferences using Oracle 11g