Explore how to deploy and optimize Kubeflow for MLOps pipelines on Arm64 architecture in a local, resource-constrained environment in this 15-minute lightning talk. Discover the practical implementation of Kubeflow's core qualities—composability, modularity, scalability, and portability—on a single host system with up to 192 CPU cores. Learn about deploying AI pipelines using Kubeflow Pipelines and serving large language models with KServe without requiring large GPUs or cloud services. Examine performance measurements across various workloads and understand how to meet service level objectives for advanced AI applications, including agentic AI, using local infrastructure. Gain insights into running comprehensive MLOps workflows in a "box" configuration while maintaining the full capabilities of the Kubeflow ecosystem.