Developed and deployed a weekly CI testing pipeline for the FlagOpen/FlagGems repository, targeting multiple hardware backends to stabilize test coverage and streamline feedback cycles. Leveraging GitHub Actions, Docker, and Python, the solution integrated both containerized and native test execution, automated GPU availability checks, and real-time Feishu status reporting. The modular YAML configuration enabled flexible backend support, while enhancements to environment setup improved reliability and security by securing API keys and refining dependency management. By expanding backend coverage and increasing observability through clearer test matrix naming, the work reduced manual QA effort and accelerated issue resolution across diverse deployment environments.