Microsoft has released a new report highlighting a new series of attacks targeting a toolkit called Kubeflow which is used for running machine learning operations on top of Kubernetes clusters.
The attacks first began in April of this year and have continued with the aim of installing a cryptocurrency miner on Kubernetes clusters that are exposed to the internet and run Kubeflow.
In a blog post, security research software engineer at the Azure Security Center, Yossi Weizman provided more details on Kubeflow and why nodes used for machine learning tasks are such an attractive target for cybercriminals, saying:
