Worked on the spring-ai repository to enhance the reliability of vector search workflows by addressing a critical bug in Milvus vector store integration. Focused on ensuring that similarity searches correctly passed the database name parameter, aligning with Milvus 2.3.5+ API requirements and preventing cross-database data retrieval. This fix improved data integrity and reduced the risk of data leakage across databases. The work involved debugging cross-database queries and maintaining compatibility within the Java and Spring ecosystem, demonstrating expertise in Java development, Spring AI, and vector databases. The contribution strengthened the accuracy and safety of search operations in production environments.