Developed a foundational AI Behavior Profiling Sample Dataset for the Guardian repository, focusing on patient information, vital signs, medication adherence, and behavioral observations. Leveraged data engineering skills to design a schema-driven structure, ensuring consistency and reproducibility across both CSV and JSON formats. The dataset serves as a core resource for training and evaluating Guardian’s AI models, supporting benchmarking and accelerating model iteration cycles. Emphasized dataset creation and schema definition, with careful attention to data modeling and artifact version control. The work established a reproducible workflow for data science teams, enabling more efficient experimentation and reliable evaluation of AI-driven behavior profiling.