Worked on the FAIR-Chem/fairchem repository to address a critical bug in ASE dataset handling, focusing on data preprocessing and database management using Python. The main contribution involved fixing a data copying issue during dataset splitting, ensuring that both row data and associated metadata were preserved throughout the process. Additionally, refined the conversion of ASE database rows into graph representations, enabling accurate extraction and attachment of extra information for downstream analytics. Emphasized robust testing to validate data integrity and maintainability of preprocessing pipelines. The work improved reliability for subsequent graph analytics workflows by safeguarding metadata during database operations and transformation steps.