Developed and delivered a security-focused feature for the opendatahub-io/kserve repository by implementing confidential model serving capabilities. This work introduced a new 'confidential' property within service configurations, enabling encrypted model artifacts and allowing users to specify decryption key resource IDs for secure deployment. Leveraging Go and YAML, the solution integrates with Kubernetes environments and aligns with cloud computing and DevOps best practices. The enhancement addresses compliance requirements for sensitive models, reducing the risk of artifact exposure and supporting secure-by-default deployment patterns. The contribution demonstrates a targeted approach to privacy and governance priorities within the model serving infrastructure.