• Developing a centralized data marketplace using React, integrating a chatbot for enhanced search capabilities across APIs, documentation, and data resources, improving accessibility and self-service for users• Integrating third-party microapps and Microsoft Fabric to streamline data management and enhance user experience

Identified over 10,000 account anomalies using an ensemble of One-Class Classification models (One-Class SVM, Isolation Forest)• Classified thousands of accounts into specific company programs by implementing multiclass XGBoost, potentially resulting in an estimated $1.6 million annual revenue increase• Managed and transformed over 150,000 database entries with SQL functions and stored procedures

• Led a data science team providing analysis and statistical modeling for nonprofits• Analyzed, preprocessed, and cleaned dataset of consisting of over 2000 clients using pandas• Produced over 15 high-quality visuals using Plotly and presented findings, providing actionable insights on client demographics and outreach strategies

• Developed binary classification Model with Vision Transformers achieving .98 f1 score in detecting malignant skin lesions• Demonstrated that obstructions on skin lesions reduce model F1 score by 45%, highlighting the need for obstruction-removal techniques• Fine-tuned VGG16, ResNet50, and MobileNetV2, experimenting with various models to optimize performance for skin lesion classification

• Collaborated with a team of interns to develop a robot car capable of simulating real-life traffic scenarios that passed 85% of driving conditions• Utilized Unreal Engine 4 to create realistic simulations of traffic conditions• Shadowed experienced mentors, gaining valuable insights into industry standards and best practices in software engineering