Yuvraj 🧢
I run 1000+ node clusters in production and send most of what I learn back upstream to Kubernetes. I like the hard parts: the scheduler, the control plane, and the point where scale turns a simple idea into a real problem.
Latest Posts
About Me
I learn by building real systems and shipping them, and open source is where most of that learning goes back. I would rather solve a problem in production than memorize the theory of it.
My edge: relentless curiosity and a bias for shipping. I learn fastest with my hands on real systems, and it shows in the work.
What I do
- Build and operate a multi-cloud Kubernetes platform (Namespace- and Cluster-as-a-Service, 1000+ nodes) for ML workloads at Woven by Toyota.
- Member of the Kubernetes GitHub organization and one of the top contributors based in Japan.
- Contribute to Kubernetes SIG projects: LWS (LeaderWorkerSet), JobSet, and DRA (Dynamic Resource Allocation).
- Earlier a core contributor to flyte.org and kyverno.io.
- Work across HPC and large-scale training, large-scale LLM inference, full-stack applications, and Kubernetes infrastructure at scale.
- Self-taught engineer who learns best by building real systems.