Worked on automation and observability improvements across Kubernetes and Python projects. In the menloresearch/litellm repository, implemented automated TTL cleanup for migration jobs within the Helm chart, leveraging YAML and Helm Charts to streamline job lifecycle management and reduce cluster resource usage. In traceloop/openllmetry, addressed reliability in Google Generative AI instrumentation by refactoring the wrapping logic in Python, ensuring correct handling of asynchronous operations within OpenTelemetry integrations. Demonstrated a methodical approach to DevOps and API instrumentation, focusing on maintainability and resource efficiency. The work emphasized production-ready changes, release discipline, and improved observability for cloud-native and AI-driven environments.