Worked on enhancing the reliability and maintainability of nested embedded document handling in the voxel51/fiftyone repository. Focused on robust merging and schema inference for deeply nested EmbeddedDocumentField structures, supporting both dict and list formats as well as subclassed fields. Refactored legacy list handling to streamline processing and reduce complexity, while strengthening test coverage with comprehensive unit and integration tests to prevent TypeErrors and catch regressions. Leveraged Python for backend development and data modeling, emphasizing clear schema inference and maintainable code. The work improved the correctness and resilience of nested schema pipelines, ensuring accurate field conversion and validation throughout the codebase.