Worked on the scalameta/scalameta repository to enhance reliability and reproducibility in the semanticdb generation pipeline. Addressed a key issue by implementing deterministic sorting for TextDocument proto files, ensuring that symbols, occurrences, diagnostics, and synthetics are consistently ordered by their ranges. This approach, using Scala and leveraging compiler plugins and code generation techniques, introduced implicit ordering for Range objects to facilitate stable sorting. The solution resolved nondeterministic outputs across compilations, reducing CI flakiness and improving the accuracy of regression testing. The work demonstrated a strong understanding of deterministic data processing and contributed to more reliable build processes in Scala tooling.