• Design and implement end-to-end ETL pipelines using PySpark and SQL, integrating structured and unstructured data into Azure Data Lake.• Reduced the runtime and cost of our largest data pipeline by 90%, resulting in weekly savings of $1K.• Act as the primary data point-person for large-scale projects, ensuring stakeholder alignment on requirements, progress, and outcomes.

• Conducted source-to-target data mapping, defining data lineage and ensuring a clear understanding of data movement throughout the organization.• Established and maintained data quality frameworks, performing ETL validation to ensure accurate and reliable information for downstream analytics.

• Leveraged Python and SQL to resolve data quality issues, procure new data sources, and provide ad-hoc analyses, ensuring timely delivery to stakeholders.• Led major effort to to estimate monthly energy consumption by implementing ML models (random forest, ARIMA), achieving 85% accuracy.• Redesigned ETL pipeline to ingest complex time series data into PostgreSQL database, improving performance by 30%.• Streamlined data verification process by automating invoice downloads with Python (Pandas, Selenium), significantly reducing manual effort.• Developed a Python program to handle client outreach, accelerating the quarterly task from a day to under an hour.

• Evaluated company’s wellness program by applying hypothesis tests to pre- and post-survey data; presented statistical results in a digestible form for a non-technical team.

• Collaborated across departments to deliver complex queries of residential/commercial real estate data consisting of ~10M+ rows.• Created a logistic regression model in SAS to predict whether homeowners would purchase a heat pump, achieving 75% accuracy.• Overhauled efforts to consolidate 20+ data sources by parsing and standardizing messy address text, improving previous match-rate by 40%.• Developed a multiple linear regression model and Excel tool (VBA, Power Query) to estimate household air leakage, allowing engineers to simulate tests during the pandemic.

• Supported research in computational molecular dynamics to determine 3D biomolecule structures.• Wrangled unstructured data for simulations by automating text-cleaning procedures with Python.