Worked on the acryldata/datahub repository to optimize the MongoDB metadata ingestion pipeline, focusing on improving performance and scalability for large data collections. The approach involved reordering aggregation stages within the pipeline to enable early sampling or dataset limiting, which reduced the volume of data processed by subsequent steps and increased ingestion throughput. Implemented a new test to validate non-random sampling behavior, ensuring correctness of the optimized path. Utilized Python for development, applying skills in data ingestion, database integration, and performance optimization. The enhancement enabled faster metadata availability for downstream analytics and improved readiness for scaling ingestion workloads.