Dmitri Gekhtman
Senior Software Engineer at NVIDIA
San Francisco, California, United States
Summary
🤩
Rockstar🎓
Top SchoolDmitri Gekhtman is a Senior Software Engineer based in San Francisco with five years of hands-on experience building compute and ML infrastructure for large-scale distributed systems. He has driven autoscaling, scheduling, and Kubernetes operator work at companies like Cruise and Anyscale, and now focuses on GPU cluster scheduling at NVIDIA. Dmitri is a practical expert in Ray and KubeRay—contributing to autoscaler, CI, and Kubernetes integrations for the widely used Ray project—and has improved deployment, testing, and default configurations that make distributed ML more reliable. With a PhD-trained mathematical background, he blends rigorous problem-solving with production-grade engineering, and he’s known for making complex resource-management features (gang scheduling, priority preemption) intuitive for ML teams.
5 years of coding experience
6 years of employment as a software developer
California Institute of TechnologyBachelor's degree Mathematics Physics secondary, Bachelor's degree Mathematics Physics secondary at Harvard University
Stackoverflow
156reputation
2kreached
5answers
1question
Github Skills (23)
autoscaling10
kubernetes10
docker10
python10
testing10
dockers10
cicd10
automation10
automation-testing10
kubernetes-pods10
ray10
automations10
devops10
scripting9
go9
Programming languages (8)
Github contributions (5)
ray-project/kuberay
May 2022 - Jan 2023
A toolkit to run Ray applications on Kubernetes
Role in this project:
DevOps & Backend Engineer
DevOps & Backend EngineerContributions:1 release, 628 reviews, 60 commits in 8 months
Contributions summary:Dmitri primarily contributed to the project's infrastructure and backend configurations, especially around the Ray autoscaler and Kubernetes integration. Their work included upstreaming changes from the Ray repository, modifying the autoscaler container's configurations, and adding autoscaler-related features to the RayCluster CRD. Further contributions focused on ensuring the correct functioning of the autoscaler by fixing a service account typo and improving default settings.
raydeep-learningapachemachine-learningkubernetes
ray-project/ray
Oct 2020 - Jan 2023
Ray is an AI compute engine. Ray consists of a core distributed runtime and a set of AI Libraries for accelerating ML workloads.
Role in this project:DevOps Engineer
DevOps EngineerContributions:1 release, 1212 reviews, 225 commits in 2 years 4 months
Contributions summary:Dmitri focused on improving continuous integration (CI) processes and build automation within the repository. Contributions include integrating type checking (mypy) into the CI workflow, updating the build process for a more robust Docker container, and improving the Kubernetes deployment tests. The user also made contributions to test infrastructure.
pythonconsistsruntimetensorflowserving
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.