Worked on the UiPath LangChain Python repository to enhance schema generation, focusing on robust handling of nullable types in Python. Developed and integrated a process that accurately extracts non-null types when schemas include optional fields, specifically addressing cases where anyOf includes null. This approach improved the reliability of schema processing and reduced integration issues for downstream components consuming these schemas. Leveraged Python and type hinting to refine the generate_schema_from_graph function, resulting in cleaner input and output schema definitions. The work demonstrated depth in schema processing, narrowing edge cases and reducing runtime errors for external input handling within the repository.