Worked on performance optimization for the Downsample plugin in the elastic/elasticsearch repository, focusing on improving aggregation speed for time-series data. Addressed inefficiencies in the previous algorithm by implementing DocIdSetIterator-assisted document selection, which bypasses non-valued documents and accelerates metric field processing. Developed comprehensive tests to validate both correctness and measurable performance gains, ensuring reliability for large TSDB workloads. Utilized Java and YAML for backend development, emphasizing performance optimization and test-driven development practices. The enhancements led to increased throughput and reduced compute costs, enabling faster analytics and more efficient operation for users managing substantial time-series datasets in Elasticsearch.