Worked on the Higgs_collaboration_B repository to enhance the HiggsML Notebook Starting Kit, focusing on improving onboarding, reproducibility, and experimental workflows. Refactored the Jupyter Notebook structure and initialization flow, enabling more stable and repeatable data science experiments. Transitioned the modeling approach from a boosted decision tree to a neural network, aiming to improve predictive performance. Enhanced data loading processes, integrated robust logging, and added image-based visualization assets to support result interpretation. Established cross-notebook analysis pipelines to standardize workflows. Utilized Python for development, applying data cleaning techniques and version control practices to ensure maintainability and facilitate collaborative experimentation within the project.