Worked on build reliability and package management for apache/hadoop and acceldata-io/spark3 over a two-month period, focusing on targeted bug fixes rather than feature development. Addressed a critical build failure in apache/hadoop by updating the mktemp dependency in package.json, restoring CI stability and ensuring successful builds through careful dependency management and build system expertise. In acceldata-io/spark3, resolved a version metadata issue in the SparkR DESCRIPTION file, improving compatibility and runtime stability for SparkR users. Utilized R programming, JSON, and package management skills to deliver precise, review-driven changes that enhanced release readiness and maintainability across both repositories.