Zrsong enhanced the opensearch-project’s neural-search and ml-commons repositories by delivering core features for SEISMIC index management, including warm-up and cache clearing, as well as sparse index ingestion and tuning. Leveraging Java and the AWS SDK, Zrsong implemented tenant-aware encryption and global resource metadata storage using DynamoDB, improving multi-tenant security and resource utilization. The work included correcting configuration paths for neural-search index thread settings, refining integration tests, and aligning dependencies to boost CI/CD reliability. These contributions deepened backend functionality, strengthened cloud integration, and improved test coverage, reflecting a thoughtful approach to scalable plugin development and robust configuration management.