Data Engineering Intern
New York, New York
Intern within PepsiCo eCommerce's Data, Technology, and Experiences (DTX) group.-Created a reusable data validation Python Package/Docker Image/Kubeflow Component to describe and validate data characteristics within any ML pipeline. This included data visualization, inferring a schema, YAML file integration of the schema and a config, anomaly detection (against the schema and drift detection), custom functions, and slack notification. The component was configurable to any AWS S3 bucket, as to be easily integrated into any pipeline, and will become a standard pre-training component for all KubeFlow pipelines throughout the ML and DS division of PepsiCo eCommerce.-Proposed and launched a Difference-in-Differences campaign to research the effect of marketing spending on online sales of many PepsiCo products in Kroger stores across the United States. Tasks included pulling online sales data from Snowflake, creating a model to rank which product and region pair had the best pre-treatment parallel trends (e.g., OLS, Difference Variance, Prior Residual Normality/KL-Div, etc.), and visualizing the efficacy of said model.