Worked on the ollionorg/DataflowTemplates-fork repository to deliver a configurable parameter that allows Bigtable Change Streams to ignore specified column families when publishing to Pub/Sub. Updated the core Java processing logic to filter out unwanted mutations at the publish boundary, reducing downstream data noise and bandwidth usage. Developed and integrated an automated test using Dataflow and integration testing techniques to ensure the ignore functionality operates as intended. This targeted enhancement improved data quality for Pub/Sub subscribers by preventing unnecessary mutations from being emitted, while the expanded test coverage reduced the risk of regressions and increased confidence in pipeline reliability.