Scheruku enhanced deployment reliability and observability for the red-hat-data-services/trustyai-service-operator repository by introducing flexible initialization and sidecar image options, updating deployment images, and implementing comprehensive job failure monitoring with automated cleanup. Leveraging Go, Kubernetes, and controller development patterns, Scheruku improved the system’s ability to track, annotate, and reconcile failed evaluation jobs, resulting in faster deployments and clearer data lineage. In the red-hat-data-services/lm-evaluation-harness repository, Scheruku resolved Kubernetes secret mount compatibility issues in Python and strengthened MLflow integration by propagating run IDs, ensuring evaluation results are accurately linked to their provenance and improving traceability across machine learning workflows.