Worked on the Firedrake repository to improve the reliability and scalability of parallel MPI simulations, focusing on the particle swarm mesh logic. Addressed a critical IndexError that occurred when MPI ranks encountered empty cell partitions by introducing guarded cell closure indices and a sentinel value, ensuring robust handling of edge cases. Enhanced the codebase using Python and parallel computing techniques, and expanded regression test coverage to maintain stability in large-scale runs. The work prevented production crashes and improved predictability in distributed numerical simulations, reflecting a strong emphasis on debugging, testing, and maintainability within high-performance scientific computing environments.