Worked on the nashtech-garage/yas repository to enhance the Recommendations Service by building robust integration tests focused on Kafka-based Change Data Capture events and product data synchronization. Leveraged Java and Spring Boot to implement end-to-end test coverage, utilizing Testcontainers to create a scalable, containerized environment with Kafka and PostgreSQL for deterministic and reproducible testing. Refined the vector repository’s similarity search logic to improve both the accuracy and performance of recommendations. The work emphasized CI/CD readiness and test reliability, resulting in a stable infrastructure that supports ongoing feature development without introducing critical bugs, while deepening expertise in integration testing and vector databases.