cdCon+GitOpsCon 2023

GitOps is very commonly used by developers. What if I told you, you could get the befits of GitOps for your end-to-end ML jobs be it: training, inference, mentoring, testing, and more. In this talk, we start by showing the core processes that consist of an ML lifecycle and introduce Kubeflow which is an open source project for making ML on Kubernetes simple, portable, and scalable. We then show how one could do all the Machine Learning processes including but not limited to running training experiments, hyperparameter tuning, end-to-end machine learning pipelines, and training jobs while leveraging the power of GitOps. Furthermore, we also show how one could deploy multiple Kubeflow components and perform the aforementioned tasks while being deployed with Argo CD. This talk essentially equips the audience to be able to use GitOps for their Machine Learning tasks. To top it off, this talk ends with a demo of deploying an end-to-end Kubeflow ML pipeline, all with the power of GitOps.