Worked on the openshift/origin repository to enhance real-time kernel latency diagnostics by developing a three-tier statistical threshold system that distinguishes between warnings and failures. Leveraged Go and Python to unify latency result parsing, consolidating multiple workflows into a single function that provides per-CPU breakdowns and detailed statistics such as maximum, average, and P99 latency. Introduced structured JSON diagnostic artifacts to improve the clarity and richness of test results, supporting more effective data analysis and statistical modeling. Expanded unit test coverage to ensure long-term reliability, reducing mean time to resolution for latency issues and enabling more proactive detection of performance regressions.