About me

Welcome to my portfolio! Here, you can discover my projects during my Data Science Master's program at the University of San Francisco and get insights into my current role as a Data Scientist at Lemonade.

Background

I hold a Master of Science in Data Science from the University of San Francisco and a Bachelor of Arts in Business Administration from Brooklyn College. I currently work as a Data Scientist at Lemonade, where I work on end-to-end design, development, and deployment of machine learning models. With a passion for all aspects of the data science lifecycle, I possess extensive experience in creating data-driven solutions across various domains using algorithms such as GLMs (Generalized Linear Models), Boosted Trees (XGBoost), and Computer Vision (PyTorch / ResNet). Here are a few project highlights:

• Developed a pricing model that uses vehicle characteristics to forecast losses, surpassing the industry benchmark by 200%. The model employs a frequency-severity approach, with expected loss determined by multiplying the predicted claim frequency and predicted claim severity, modeled separately using Poisson and Gamma regressors.

• I developed a computer vision model to detect fraudulent vehicle photos, which achieved an average PR AUC score of 90% for underwriting purposes. Furthermore, I created a web application using Streamlit that allows stakeholders to view and interact with the model's predictions, and improve its performance through photo labeling. The Streamlit app is currently deployed on an AWS EC2 instance.

• Migrated real-time machine learning models from legacy inference infrastructure to AWS SageMaker. Additionally, I built an automated re-training and model monitoring framework using MLflow in Databricks.

• Managed, ran, and presented critical SOX and internal controls to external auditors quarterly, built off telematic data. Ultimately fully automated the controls using Snowflake and Tableau dashboards.