Experienced software engineer, author, and technical lead specializing in infrastructure and backend systems with a proven track record of building and contributing to large-scale projects.
Languages: Go, Typescript, Java, Python, Rust
Systems: Kubernetes, containers, Linux, systemd, Terraform, AWS
Production Kubernetes is a 500+ page book covering a variety of topics related to building platforms and distributed computing. It is the amalgamation of years of experience architecting, deploying, and extending Kubernetes-based platforms.
I left VMware to try my hand at co-founding Arctir, where I'm focusing on defining engineering practices and building the platform. We are an early (pre-revenue) company building a platform that improves visibility into an organization's software landscape. Our stack includes Go, Typescript, React, AWS, and Postgres.
I joined VMware as part of the Heptio acquisition. At VMware, I built the technology and team that was VMware Tanzu CE, VMware's open-source Kubernetes distribution. This grew into one of VMware's largest open-source projects, deployed in a variety of scenarios from production platforms to local development environments. As the tech lead, my work included creating design proposals, implementing (coding) features, growing the technical aptitude of the team, and interfacing with product management to help shape the roadmap.
I joined Heptio, founded by 2 of the creators of Kubernetes, as an early engineer focusing on field engineering. In Field engineering we are embedded engineers within strategic accounts and help them architecture, deploy, scale, and extend their Kubernetes deployments. During my time in this role, I architected and deployed Kubernetes at some of the largest banks and telcos in the US. Along with architecture work, development was often required to make Kubernetes fit within requirements. This often required writing controllers (typically in Go, sometimes Python), which could extend functionality in Kubernetes. Multiple implementations done at customers were brought back and built into product, such as our namespace provisioner, which enabled self-service environments to teams in an organization. In my work at Heptio, enablement was a huge focus and I have helped many SRE teams get the skill sets needed to support their newly built platform over time. Much of the insight gained from my experience at Heptio was translated into my book Production Kubernetes (O'Reilly). We (Heptio) eventually exited via an acquisition by VMware.
I joined CoreOS as I was excited about their many projects at the forefront of containers and Kubernetes. Including but not limited to etcd, rkt, and fleet. I was an early engineer brought in to split my time between product development and acting as a resident engineer with our customers. While I was involved in a variety of customer deployments, our most notable was with Ticketmaster where we were one of the earliest shops to use Kubernetes in production. Getting to production involved a lot of engineering and platform extensions to make it meet requirements, namely around ingress and network-layer packet routing. This project was a highly-celebrated success along with being one of CoreOS's largest revenue streams. With my time at CoreOS I also authored the AWS Ingress controller, which is used in Amazon EKS today, and built multiple features for Tectonic, CoreOS's Kubernetes platform, which much of the technology was used to power OpenShift. We (CoreOS) eventually exited via an acquisition by RedHat.
I joined MuleSoft with an intent to transition from technical training to full-time software engineering, which I did. This was my first exposure to software at massive scale. My feature work was writing the Java-based API gateway product that was critical path for all API requests used in one of the world's largest food chains. It featured a world-wide cut over and request-per second load that is still one of the highest of my career. Along with software engineering, I facilitated multiple training courses across the globe that enabled Java developers to build integrations with the Mule framework. We (MuleSoft) IPO'd and upon becoming public I transitioned out to CoreOS.
Joining Hyland Software was my first foray into software development outside of personal projects. I built curriculum and taught software standards for technical internal employees, often onboarding software engineers, such that they could be brought up to speed with the technical stack of Hyland. During this time I fell in love with software engineering and knew I wanted to find a company open to providing me the opportunity to move out of training into engineering. This motivation brought me to MuleSoft.
Spent a few years in Apple retail software and hardware training in the ecosystem. My work here sparked the initial fire in my love for Unix and desire to move into software development full time.
I recommend visiting my website, joshrosso.com, for a full understanding of my many projects and talks. However, below is a list of some notable ones.