Worked on the duckdb/duckdb-httpfs repository to enhance the reliability of S3-backed data ingestion by addressing a critical bug in the S3 COPY workflow. Focused on correcting the handling of FILE_SIZE_BYTES, the work ensured that CSV file size reporting accurately reflects actual file sizes, which is essential for downstream ETL processes and analytics. Utilized C++ and SQL to update both the core logic and associated tests, improving test coverage and data integrity. The changes reduced the risk of incorrect file-size accounting, supporting more accurate cost and performance analysis for users relying on S3 storage and database management workflows.