At Home Depot Canada, I build and evaluate production GenAI systems, including automated QA infrastructure for a complex 40+ agent system on Google ADK. I created DeepEval and pytest test suites with Streamlit diagnostics, helping identify and fix faulty agent logic across most agents.
I also re-architected a geospatial Airflow pipeline into an on-demand notebook workflow, reducing recurring Mapbox costs by approximately 97%. My work includes a modular alert-ranking system that prioritizes the most business-relevant signals and creates a foundation for future recommendations.
Previously at OMD Canada, I developed forecasting, marketing mix modeling, and audience-segmentation solutions using Random Forest and XGBoost to inform targeting decisions. I led a Snowflake-to-Amazon Redshift migration in two months, improving performance while reducing infrastructure costs.
Across analytics, machine learning, and GenAI, I translate ambiguous business problems into practical technical solutions and collaborate closely with Data Science, Engineering, Product, QA, and DevOps teams.
