Worked on the Subaru-PFS/datamodel repository, focusing on targeted improvements to stability and installation reliability. Addressed a critical compatibility issue by updating PfsConfig to use np.nan instead of the deprecated np.NaN, ensuring correct nanmean calculations and seamless operation with NumPy 2.x. Later, streamlined the installation process by removing an obsolete ICS workaround from the shell script, reducing setup complexity and eliminating a known failure point. These contributions emphasized careful, well-scoped changes using Python, NumPy, and shell scripting, resulting in a more robust codebase with reduced upgrade risk and improved readiness for broader deployment without introducing unnecessary complexity.