During July 2025, Will focused on enhancing reliability in the kubernetes/kops repository by addressing a concurrency issue in GCE-backed cluster provisioning. He implemented logic in Go to ensure that disk creation operations on Google Cloud Platform completed synchronously before subsequent provisioning steps, using cloud API handling and infrastructure as code principles. This change prevented race conditions and reduced the risk of clusters entering a partially provisioned state. Will’s work demonstrated a clear understanding of cloud infrastructure workflows and concurrency management, resulting in a targeted, maintainable fix that improved the robustness of GCP-based provisioning within the kops codebase.