You are a Machine Learning Engineer at a tech company that has developed a model using AI Platform. The model is now being moved to production, where it serves a few thousand queries per second. However, you're encountering latency issues. The current setup involves a load balancer distributing requests across multiple Kubeflow CPU-only pods on Google Kubernetes Engine (GKE). Given the constraints of not altering the underlying infrastructure, which of the following strategies would be the MOST effective in improving the serving latency? (Choose one correct option) | Google Professional Machine Learning Engineer Quiz