Kubernetes has revolutionized cloud infrastructure by supporting scalable, containerized applications, and its utility has extended to AI/ML workloads, posing challenges such as specialized hardware needs and complex data pipelines. Google Cloud Kubernetes Engine (GKE) offers a strong foundation for these workloads, but manual infrastructure management can be cumbersome, prompting the use of Pulumi, an Infrastructure as Code (IaC) tool that automates and simplifies AI/ML infrastructure management on Kubernetes. Pulumi enables AI/ML teams to use familiar programming languages like Python to provision and scale Kubernetes clusters, integrate infrastructure with machine learning pipelines, and automate deployments, thereby reducing operational overhead. For use cases such as deploying large language models with Retrieval Augmented Generation or training and serving custom models, Pulumi allows for automated, scalable, and secure management of the entire AI/ML lifecycle, enhancing scalability, cost efficiency, and security while ensuring compliance through policy-as-code practices.