Open Source
My open-source work has been focused on platform engineering and AI.
Foyle
Foyle is a project I created to use AI to simplify deploying and operating software.
Kubeflow
I created Kubeflow while I was at Google working on what is now Vertex AI. Kubeflow started as a way to build what was then Cloud ML Engine on Kubernetes by creating a custom resource for TensorFlow Jobs. We ended up releasing that as OSS. This project morphed into Kubeflow which we announced at Kubecon EU in 2018.
Hydros
Hydros started as a project at Primer to automate hydrating Kubernetes manifests. I have continued to develop it as a way to do continuous delivery in a declarative, level based way. I wrote a blog, Data Science Meets Devops: MLOps with Jupyter, Git, & Kubernetes, that describes the pattern in more detail.