Yannis Zarkadas
Software Engineer at Google
United States
Summary
🤩
Rockstar🎓
Top SchoolYannis Zarkadas is a software engineer with nine years of experience, currently optimizing performance and correctness for the XLA compiler at Google. He holds a CS PhD from Columbia and brings deep systems expertise from work on eBPF, TPU performance, and kernel scheduling during Google internships. Prior roles at Arrikto and sustained open-source contributions to high-profile Kubernetes projects like Kubeflow and the ScyllaDB operator show strong DevOps and cloud-native deployment pedigree. His contributions often focus on build/release automation and reliable storage/configuration for ML platforms, reflecting a pragmatic blend of research and production engineering. Comfortable across compilers, distributed systems, and Kubernetes, he combines low-level optimization skills with practical infrastructure improvements. Outside the obvious academic background, he consistently bridges research insights into tangible improvements in large open-source ML tooling.
9 years of coding experience
3 years of employment as a software developer
Minor Finance, Minor Finance at The American College of GreeceDiploma Electrical and Computer Engineering, Diploma Electrical and Computer Engineering at National Technical University of AthensHigh School, High School at Arsakeio School in Psychiko
English, Greek, German
Stackoverflow
1reputation
0reached
0answers
0questions
Github Skills (18)
kubernetes10
docker10
kustomize10
makefile10
dockers10
cicd10
gcp10
release-management10
automation10
go10
kubernetes-pods10
yaml10
automations10
operator10
bash9
Programming languages (13)
Github contributions (5)
kubeflow/kubeflow
May 2019 - Jun 2021
Machine Learning Toolkit for Kubernetes
Role in this project:
DevOps Engineer & Release Manager
DevOps Engineer & Release ManagerContributions:2 releases, 42 reviews, 57 commits in 2 years 1 month
Contributions summary:Yannis primarily contributed to improving the build and release process for the Kubeflow project. Their work involved fixing and refactoring build targets within Makefiles, addressing issues with Dockerfile builds using BUILDKIT, and implementing a mechanism for specifying a static configuration file. Moreover, the user added features to the `kfctl` tool, including a `--config` flag and a mechanism to emit image changes to standard output for better release management. These changes streamline the build and deployment processes for the Kubeflow toolkit.
pythondata-sciencenotebookmachine-learningmlops
kubeflow/manifests
Oct 2019 - Jun 2021
Kubeflow Deployment Manifests
Role in this project:DevOps Engineer
DevOps EngineerContributions:1 release, 9 reviews, 46 commits in 1 year 8 months
Contributions summary:Yannis's commits primarily focus on modifying and adding configurations for persistent volume claims (PVCs) within the Kubeflow manifests. They've added default storage classes, modified file names, and edited code related to persistent volume claims in the minio and mysql pipeline components. Furthermore, they've worked on parameterizing and adding missing images to kustomizations, indicating involvement in build and deployment processes. The changes suggest an effort to improve the deployment configuration and infrastructure management of the Kubeflow platform.
kubeflowk8skuberneteskustomizemanifests
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.