Worked on stabilization and performance improvements for Iceberg-backed scans in the linkedin/openhouse repository, focusing on a critical bug fix in the scan optimizer. Addressed issues with datetime filter handling and CAST expressions, restoring partition pushdown for date and time predicates in partitioned Iceberg tables. Utilized Python for both the core logic and comprehensive test coverage, ensuring that conversion functions correctly process TIMESTAMP, DATE, and TIME types. Validated the solution end-to-end, confirming that daily-partitioned scans now efficiently prune manifests based on datetime predicates. Demonstrated strong skills in Python programming, data engineering, and database management through targeted, high-impact backend improvements.