Developed a data preparation pipeline for the apache/singa repository, focusing on malaria image datasets used in the healthcare model zoo. The work involved building a Python-based data loader and preprocessing script that loads, resizes, normalizes, and validates the existence of training and testing images, streamlining the onboarding of new datasets. By standardizing data storage and preparation, the solution improved data readiness for machine learning experiments and reduced setup time and errors. Leveraging skills in data loading, image preprocessing, and machine learning data preparation, the developer established an end-to-end flow that aligns with model zoo requirements for reliable model training.