Worked on the apache/druid repository to address data aggregation accuracy during index merging, focusing on the FixedBucketsHistogram component. Applied Java skills to implement a reset mechanism within the AggregateCombiner, resolving over-counting issues and ensuring histogram integrity. Enhanced unit testing by specifying character encoding for byte retrieval, which stabilized test outcomes and reduced false negatives. The work emphasized maintainable, traceable changes and demonstrated a disciplined approach to debugging and code quality. By improving the reliability of histogram-based insights during index merges, the contributions supported more accurate query results and reinforced the overall robustness of data processing within the project.