💹 About Me

Data Scientist with 10+ years of experience using advanced analytics and predictive modeling to solve complex, ambiguous problems. Proven leader in translating vast datasets into actionable, high-impact business strategy for executive leadership.

Expert in: econometric modeling, causal inference, and a full suite of data tools including Python, R, and SQL.

<aside> 🎯

Quick Stats

✅ 10+ years experience ✅ Led forecasts for global energy and agriculture markets ✅ Published in The World Economy ✅ Multiple U.S. government publications

</aside>

<aside> 💻

Explore All Code & Methodologies

Relevant Code to view detailed implementations, scripts, and technical documentation for all projects featured in this portfolio.

</aside>

<aside> 📚

View All Publications

Publications to explore peer-reviewed research, government reports, and technical articles.

</aside>

<aside> 🏆

View All Awards

Awards to explore professional recognition, honors, and achievements.

</aside>

🚀 Featured Projects

Project 1: Production Time-Series MLOps: Forecasting Global Energy Futures ⚡🔋🛢️

<aside> 🎯

The Challenge: As lead production engineer for the Short-Term Energy Outlook (STEO), I owned the end-to-end predictive modeling infrastructure for Europe and Eurasia petroleum and LNG supply forecasting—a mission-critical system directly informing billions of dollars in private sector investment and national energy policy decisions. The system required real-time accuracy, scalability, and reliability under extreme stakeholder scrutiny.

</aside>

The Solution: I engineered end-to-end analytical solutions using Python and R, deploying production-grade time-series models (ARIMA/Prophet) within a robust MLOps framework. Core technical components included:

• Automated ETL/DE Pipeline: Built and maintained a production ingestion framework that orchestrated multi-source data acquisition, validation, transformation, and feature engineering for time-series inputs—ensuring data quality and pipeline reliability at scale.
• Time-Series MLOps: Developed, validated, and deployed advanced predictive modeling systems for petroleum and LNG supply, implementing continuous monitoring, retraining workflows, and A/B testing protocols to ensure forecast accuracy and model drift detection.
• Deployment & Observability: Integrated model outputs into Power BI dashboards for executive consumption, providing real-time forecast visualization, confidence intervals, and scenario analysis for strategic decision-making.

<aside> 💻

📋 Code & Methodology Reference: View the advanced time-series and econometric methods used in this project

</aside>

<aside> ✨

Business Impact: The production systems I built and maintained became a foundational component of one of the world's most-watched economic forecasts. The predictive models directly informed billions of dollars in annual private sector investment, guided U.S. energy policy at the executive level, and maintained 99%+ uptime in a high-stakes production environment.

</aside>

Project 2: Causal Inference & Consumer Segmentation Under Exogenous Risk Events (Master's Thesis) 🎓📊⚾

<aside> 📋

View Full Presentation: Master's Thesis Presentation (Google Slides)

💻 Full Stata Code: See Relevant Code → Causal Inference System: Terror Alerts and Consumer Behavioral Response

</aside>

<aside> 🎯

The Problem: Product teams and risk analysts need to quantify how exogenous threat events impact real consumer behavior across demographically diverse user segments. Traditional approaches relied on survey data and self-reported sentiment, which are noisy, biased, and non-actionable for business decision-making. The challenge was to measure actual observed behavioral shifts in response to external risk signals and identify which customer segments exhibited the strongest response—a critical capability for product attribution, targeted marketing, and business continuity planning.

</aside>