Worked on the autowarefoundation/autoware.core repository, focusing on enhancing reliability and correctness in point-cloud processing for autonomous vehicle perception. Addressed a logic error in the voxel grid downsampling filter by refining the finite-point check, which ensured accurate downsampling and prevented invalid data from affecting downstream components. Utilized C++ for development, applying skills in algorithm optimization and data processing to align with repository quality standards. Documented the changes thoroughly to support traceability and future maintenance. This targeted bug fix improved the stability of the perception pipeline, ensuring that voxel-based downsampling operated as intended and maintained data integrity throughout the workflow.