In November 2025, Ada Wrapub focused on backend development and data engineering within the apache/iceberg repository, addressing a complex bug in Puffin Deletion Vector handling. Ada implemented a targeted fix to ensure manifest and Puffin blob metadata accurately reference data file paths, preventing failures in the deletion vector reader during delete operations. Using Java and Apache Spark, Ada introduced utility methods for updating metadata and preserving bitmap data during path rewrites. The solution was applied across both the core data model and Spark execution paths, reinforcing data integrity and stability in deletion workflows for the Puffin format.