Jonathan Tong smiling in front of a brick wall

Hi, I’mJonathan.

I build open-source AI and Kubernetes infrastructure, turning complex distributed systems into usable tools.

Jonathan Tong standing on a New York City street

About me.

I’m a software engineer building open-source AI, Kubernetes, and cloud infrastructure. I enjoy turning complex systems into reliable tools that are easier for people to use.

Education

Cornell University

B.S. in Computer Science

College of Engineering Minor in Information Science May 2021

Based in

New York City

Originally from

Clemson, SC

Current work.

Microsoft software engineering across open-source AI, Kubernetes, and Azure, followed by previous roles in cloud infrastructure and machine learning.

Current role —Present

MicrosoftOpen source · AI · Kubernetes · Azure

Software Engineer II

I work on open source projects in AI and the Kubernetes ecosystem. My focus is on project design, implementing new features, cross-company launches, project governance, and contributor experience. I bring the changes from upstream into production environments on Azure.

  • AI infrastructure
  • Kubernetes upstream
  • Azure productization

AI + Kubernetes

I use my expertise in open source project governance and Kubernetes development to cutting edge AI projects.

  • Helped transition KAITO from an internal AKS codebase to a cloud-agnostic open-source project.
  • Set release and docs conventions and redesigned BYO GPU nodes for any provider or SKU.
  • Built zero-config AKS setup for GPU drivers and RDMA/InfiniBand.
  • Added training and inference validation and proposed a one-click cluster-to-Dynamo path.

Kubernetes Upstream.

I build and maintain projects across several Kubernetes SIGs, focusing on test reliability, API evolution, governance, and contributor pathways.

kubernetes-sigs/cluster-api
clusterctl reviewer

Authored the add-on orchestration proposal and implemented MachinePool Machines across core CAPI and CAPZ.

Jont828/cluster-api-visualizer
Creator & maintainer

Built an interactive resource-tree UI for debugging Cluster API clusters; reached 65K downloads and added v1beta2 conditions support.

Azure Integration.

I connect public Kubernetes projects with Microsoft’s build, validation, artifact, and fleet-management systems so upstream development can be consumed through repeatable production paths.

Fleet platformCAPI · Argo CD

Integrated Cluster API with GitOps workflows to install cluster components and support a fleet of more than 50 clusters, including web UI support.

Dynamo Qwen3-235B · Kubernetes Demo preview Show playback controls
Live deployment demo Dynamo profiles, configures, and brings a distributed Qwen3-235B inference graph online on Kubernetes.
One resource tree connects core Cluster API, Azure infrastructure, machine pools, and declarative add-ons.
Upstream changes move through reproducible builds and end-to-end validation before reaching Azure consumers.

Previous roles.

Cloud networking, storage systems, and applied machine learning shaped the systems perspective behind my work today.

Microsoft Azure Networking

Software Engineering Intern

  • Built an experimental circuit-breaking feature for Open Service Mesh to limit network traffic between services.
  • Designed a Kubernetes custom resource definition for circuit-breaking settings such as maximum connections and requests.
  • Watched custom-resource changes in Go and configured Envoy through xDS so updates applied in real time.
  • Added concurrent-request and request-rate configuration to the Open Service Mesh demo.
  • Contributed the work as an experimental Service Mesh Interface extension for the CNCF community.

KubernetesPolicy change Envoy xDSLive proxy update

Oracle Cloud Infrastructure

Software Engineering Intern

  • Created Java and Python command-line tools to simulate volume placement in production regions, output statistics, and graph the results.
  • Identified issues in the volume-placement algorithm, proposed changes, and used the simulation tools to measure the improvement.
  • Applied the simulation findings to improve host-server selection for the largest volume class.
115% Placement capacitymore 32 TiB volumes placed in simulation

Clemson University Computer Science Research

Research Intern

  • Developed a pronunciation-based convolutional neural network for cyberbullying detection under computer science professor Feng Luo and co-authored two research papers.
  • Used Python and Bash to run machine-learning models including Naive Bayes, Random Forests, and support vector machines.
  • Converted text to a phonetic encoding to reduce noise from internet slang and misspellings, producing measurable gains across precision, recall, and accuracy.
+8.6%Precision +8.9%Recall +6.3%Accuracy

My publications.

Two peer-reviewed papers grew from the Clemson research program, connecting language, security, and machine learning.

Skills & tools.

A concise inventory of systems, languages, and methods spanning infrastructure, tooling, and research.

Systems & infrastructure

  • Distributed systems
  • Kubernetes operators
  • AI inference infrastructure
  • GPU orchestration
  • Cluster lifecycle
  • GitOps
  • Linux

Languages

  • Go
  • Java
  • Python
  • C
  • OCaml
  • Bash

Cloud & open source

  • Microsoft Azure
  • NVIDIA Dynamo
  • KAITO
  • Cluster API
  • CAPZ
  • CAAPH
  • CNCF ecosystem

Research & developer tools

  • API design
  • Controller development
  • End-to-end testing
  • Release engineering
  • Developer documentation
  • Community leadership
  • Web interfaces

Let’s get in touch.

For engineering conversations, open-source collaboration, or research connections, pick a favorite and say hello.