During March 2026, contributed to the castai/k8s-agent repository by developing a Kubernetes Percentage-Based Resource Allocation feature. This enhancement allowed resource requests and limits to be specified as percentages, improving resource utilization and scheduling accuracy in heterogeneous clusters. The work involved updating core data structures to support percentage-based allocation, requiring a disciplined approach to backend development and a strong understanding of Kubernetes resource management. Implemented primarily in Go, the feature enables more predictable deployments and cost optimization for users. No major bug fixes were recorded during this period, with efforts focused on delivering robust, feature-driven improvements to the codebase.