Used exploratory data analysis in SQL and Python (Azure/Databricks) to identify a recurring inconsistency, automating a 10-step manual process to 2 steps (80% reduction in manual effort). Practiced transparency by flagging a data quality issue at its root rather than letting a report quietly stay wrong, documenting the fix from scratch in Confluence, still in use today. Led with an AI-first mindset, using Claude and GitHub Copilot daily to accelerate development, and took the initiative to complete dbt Fundamentals independently.

Designed and validated the methodology behind 5 machine learning models (Random Forest, XGBoost, SVM, KNN Regressor), applying rigorous exploratory data analysis to multi-source data. Improved model accuracy and R 2 by 15%+ versus baseline through rigorous experimental design and cross-validation. Delivered honest, transparent data storytelling to non-technical research leads and international sponsors, always upfront about the limitations of what the data could confirm.

Built automated SQL/Tableau dashboards, reducing manual corporate reporting by 60%. Translated business questions into clear, actionable insights for stakeholders across teams.

Built a customer classification model to predict purchase volume, improving marketing targeting precision by 25% through statistical testing and data validation.