Contributed to climate modeling and high-performance computing projects by enhancing documentation and stabilizing GPU workflows. Improved the NCAR/musica repository’s tutorial on user-defined reactions, clarifying that reaction rates are provided directly and refining guidance for ambiguous observational scenarios. This work, implemented in Jupyter Notebook and Fortran, strengthened onboarding and reproducibility for new users. In the ESCOMP/atmospheric_physics repository, addressed a broken GPU test in the CAM model’s RRTMGP gas optics calculation by removing unnecessary memory allocation and directly accessing mass mixing ratios, which improved test reliability without impacting performance. Demonstrated technical writing, Fortran, and GPU computing skills across both projects.