Worked on enhancing reliability and resource management in the EnzymeAD/Reactant.jl repository by addressing a memory leak in the to_rarray context management. Focused on memory management and performance optimization using Julia, the solution involved ensuring that MLIR contexts created during repeated to_rarray executions were properly disposed of after use. This approach prevented incremental memory growth and reduced the risk of out-of-memory errors in long-running data transformation workflows. The fix was validated through continuous integration, supporting the stability and scalability of analyze and transpile pipelines. The work demonstrated a methodical approach to functional programming and maintainability in complex data processing systems.