Worked on the linkedin/openhouse repository to deliver two core backend features over two months, focusing on schema evolution and data ingestion reliability. Implemented Iceberg MergeSchema support by upgrading dependencies and aligning tests for compatibility with Spark 3.5 and Java 8, enabling safer schema changes in production pipelines. Developed write schema casing normalization across the catalog layer, allowing case-insensitive writes while preserving original table schemas and supporting multiple engines. Utilized Java, Gradle, and Spark, with extensive unit and end-to-end testing to ensure robust validation and CI reliability. The work improved ingestion flexibility and reduced schema-related errors across heterogeneous data sources.