Worked on the tesseract-ocr/tesseract repository to address a bug in ALTO XML ID generation for multi-page documents, focusing on improving the uniqueness and stability of element IDs. The solution involved refactoring the ID generation logic in C++ to conditionally include page numbers for illustrations, graphical elements, composed blocks, text blocks, text lines, and strings, ensuring that IDs remain stable on the first page and unique on subsequent pages. This change enhanced the correctness of ALTO XML output, reduced downstream processing errors, and minimized manual debugging. Demonstrated skills in C++, OCR, and XML data formatting through targeted, maintainable code improvements.