Worked on cross-repository bug fixes in Intel-tensorflow/xla and Intel-tensorflow/tensorflow, focusing on literal buffer sizing and tuple leaf array layout normalization. Addressed inconsistencies by canonicalizing element sizes within Literal::SetShape, ensuring that buffer allocation and typed materialization maintained invariant states. Used C++ and applied algorithm design principles to implement these changes, adding regression tests to validate the fixes and aligning updates with upstream imports for consistency. This work improved the stability and predictability of literal layout handling in critical compute paths, reducing the risk of incorrect buffer sizing and enhancing the reliability of software testing and development workflows.