Data Science

Experience

Principal Data Scientist — UnitedHealth Group

2024–2025

• Led evaluation of clinic optimization initiatives and partnered with IBM on a proof of concept.
• Designed A/B testing approach to measure operational changes and behavioral effects.
• Wrote internal documentation on machine learning workflows, evaluation, and Python practices.

Senior Data Scientist — 1904Labs

2022–2024

• Built an LLM evaluation tool to benchmark intent classification models using Python, Streamlit, and AWS Bedrock.
• Developed fraud and outlier detection methods for opioid prescribing patterns.
• Delivered client-facing analyses and technical prototypes for public-sector and healthcare use cases.

Senior Data Scientist — MedeAnalytics

2021–2022

• Engineered features for an infant mortality prediction model using Medicaid claims data.
• Analyzed maternal and neonatal health records to identify risk factors and intervention points.

Data Visualization Developer — Centene

2019–2021

• Built interactive dashboards for fraud detection and business planning using Plotly Dash and R Shiny.
• Deployed applications in containerized environments for internal stakeholders.

Data Scientist — EPSi

2016–2019

• Developed financial forecasting models for healthcare systems.
• Presented results to executives and translated statistical outputs into operational decisions.

Data Analyst / Developer — Booz Allen Hamilton (NASA)

2014–2016

• Co-developed NASA’s Project Cost Estimating Capability (PCEC).
• Contributed to cost modeling tools used in mission planning and budgeting.
• Received NASA awards for software innovation.

Selected Work

• Healthcare Fraud Detection: Identified anomalous opioid prescribing behavior using statistical and ML methods.
• Infant Mortality Modeling: Built predictive features from linked maternal and neonatal records.
• LLM Evaluation: Designed prompt-based evaluation frameworks for mathematical reasoning and classification tasks.
• Cost Estimation (NASA): Developed models for large-scale project cost forecasting.

Tools and Methods

• Python (Pandas, scikit-learn), SQL, R
• Jupyter, Streamlit, Plotly Dash, R Shiny
• Machine learning, statistical modeling, experimental design
• Data pipelines, feature engineering, model evaluation