Responsibilities include:

• Conducted research and methodology design for the 2030 Census Coverage Estimation program. Lead data-driven projects within the Decennial Statistical Studies Division that applied statistical and machine learning techniques to evaluate and improve the quality of the decennial census.
• Developed tract-level person and living-quarters population estimates for coverage estimation by leveraging administrative records; designed and implemented anomaly-detection and quality control frameworks; applied graph-theoretic models leveraging household structures to improve estimation accuracy.
• Advanced intercensal population estimation through the Continuous Count Study by integrating Census, commercial, and government-wide datasets. Applied log-linear and latent class modeling for characteristic imputation when implementing multiple-systems estimation. Presented findings at the 2024 Joint Statistical Meetings and the 2024 Federal Committee on Statistical Methodology.
• Provided statistical programming for the 2020 Post-Enumeration Survey (PES). Developed the in-mover probability imputation model using large-scale feature selection techniques to identify key covariates for estimation.

Responsibilities include:

• Partnered with Risk Management, Compliance, and Product teams to create automated reports and dashboards, providing insights on KPIs and OKRs using statistical modeling and data science techniques.
• Delivered high-impact analytical summaries to senior leadership, developing flexible reporting solutions to drive strategic decision-making and monitor performance indicators.
• Designed time series forecasting models leveraging public data to predict quarterly mortgage loan origination volume, optimizing workforce allocation and reducing operational costs.
• Applied Natural Language Processing to analyze mortgage process documentation, uncovering bottlenecks and reducing closing times through machine learning-based workflow improvements.