Worked on the confidential-containers/cloud-api-adaptor and kata-containers/kata-containers repositories, focusing on backend reliability and resource management. Developed a unified cleanup mechanism for orphaned virtual machines across multiple cloud providers, using Go and resource cleanup strategies to reduce costs and improve deployment hygiene. Enhanced error handling and generalized cleanup workflows to ensure robust multi-provider support. Addressed race conditions in instance deletion with mutex protection and improved deletion status reporting. Expanded runtime observability by implementing -info flag support in both Go and Rust shims, exposing runtime information for containerd v2.0+ compatibility. Maintained code quality through signed-off commits and cross-repo collaboration.