Worked on backend development and database internals to address data integrity issues in HBase, focusing on the StoreScanner component within both the apache/hbase and HubSpot/hbase repositories. Using Java, implemented fixes to the flush logic that previously allowed incorrect rows to be returned when delete cells were present, specifically by refining the detection of changes to the top cell during flush operations. Developed and integrated regression tests to ensure consistent behavior across repositories, validating that scan results remain accurate after flushes. These changes improved reliability in high-throughput scanning scenarios and reduced the risk of cross-row data leakage in production environments.