Comparison is the thief of joy. I would rather build than compete.
I build and operate large Kubernetes platforms, the kind that run 1000+ nodes in production and still have to feel simple to the teams on top of them. I am self-taught, I learn by building real systems, and I contribute most of what I learn back upstream to Kubernetes. I am happiest on hard infrastructure problems where the design actually matters, and I am open to talking with teams working on them. This page is the short version of how I got here and what I work on.
Where it started
I grew up in a big family with modest means, in the mountains, far from any tech hub. I had more curiosity than resources, so I leaned on the curiosity.
In late 2012 I found the internet and realised how much changed once you could build something and publish it for the whole world to reach. I have been learning every day since. Curiosity about building websites turned into a longer interest in infrastructure, distributed systems, and open source.
Woven by Toyota
At Woven by Toyota I build and operate a multi-tenant, multi-cloud Kubernetes platform. It offers Namespace-as-a-Service and Cluster-as-a-Service, and it powers Woven's ML workloads. In practice that means running 1000+ node clusters in production and keeping them simple to use for the teams that depend on them.
Operating at that scale keeps me honest. Every design decision carries real weight, which is the kind of problem I like.
intervyo.ai
On the side I built intervyo.ai, and my partner runs it. Together we have grown it into a profitable product. I handle the engineering, they drive the business. Two people, one team, shipping something people use.
Open source
I learn best by building real systems and sending the work upstream. Some of it:
- Kubernetes: a member of the Kubernetes GitHub organization and one of the top contributors based in Japan, working on SIG projects including LWS (LeaderWorkerSet), JobSet, and DRA (Dynamic Resource Allocation).
- Earlier work: core contributor to Flyte and Kyverno, with contributions to Falco.
- At scale: ML pipelines and distributed systems built to hold up under load.
- Tools: Go, TypeScript, and Python, and the judgment to know when to reach for which.
I understand systems the way I do because I learned them from the ground up, and that is still the part of the work I enjoy most.
How I think about the work
I measure output by impact, not by money alone. Living in the mountains, and now in Japan, taught me that the work that lasts is the work that matters to someone: building useful things, contributing to open source, learning continuously, and helping other people get further in their own careers.
What I am looking for
I do my best work close to the hard part of a system: the scheduler, the control plane, the place where scale turns a simple idea into a real problem. I want a team that is building infrastructure or ML platforms at a scale where those decisions carry weight, and that cares about getting them right. If that sounds like your team, I would like to hear about it.
The problems I want to spend the next years on:
- Kubernetes infrastructure at scale, and the distributed-systems problems that only show up there.
- HPC and large-scale training, keeping thousands of accelerators busy and coordinated.
- Large-scale LLM inference, the serving and scheduling side of running models in production.
- ML infrastructure as it grows.
- Full-stack application work, from the platform underneath to the product on top.
- Self-driving and the systems that make autonomy possible.
- Voice AI and how people will talk to computers, LiveKit in particular.
- The next generation of developer tools.
Let's connect
If you are hiring for hard infrastructure or platform work, want to collaborate on open source, compare notes on voice AI, or kick around a startup idea, reach out. I read every message, and I am always glad to hear from other people who build.