Senior Systems Software Engineer, Kubernetes Scale - DGX Cloud

at Nvidia

📍 Germany

PLN 292,500-650,000 per year

SENIOR

✅ Remote

Tech Stack
Tag name is followed by "@" symbol and proficiency level value. About proficiency levels:

1-2 — basic awareness. Minimal hands-on experience, and a rudimentary understanding of the technology's purpose;

3-6 — daily use. Comfortable and regular usage, capable of handling common tasks and challenges related to the technology;

7-9 — you are an expert, you can teach others, you know all the pitfalls and tricks;

10 — exceptional knowledge, comprehensive understanding, and adeptness in all aspects of the technology, including advanced problem-solving. Think twice before claiming or demanding such level.

AI @ 6 AWS Azure CI/CD Distributed Systems @ 7 GCP GPU Go Kubernetes @ 7 Networking @ 7 Python @ 6

Details

Role overview

The DGX Cloud organization at NVIDIA brings together cutting-edge hardware and software innovation to deliver industry-leading accelerated computing for the world’s most adventurous AI workloads. We are looking for an outstanding Senior Systems Software Engineer with deep experience in distributed systems, open-source technologies such as Kubernetes and containers, and a strong background in systems performance and scalability.

In this pivotal role, you will help scale AI infrastructure while optimizing total cost of ownership, driving down cost per token to unlock the next generation of AI innovation and AI factories.

Responsibilities

Drive end-to-end performance and scale characterization for the NVIDIA DGX Cloud software stack, from Kubernetes control and data planes through NVIDIA components such as GPU Operator, Network Operator, DCGM, NIM, and distributed inference serving, following issues from orchestration down to the metal.
Collaborate with AI researchers, developers and customers to develop innovative, automated tests that simulate real user workloads using custom-built and leading open-source tools and frameworks.
Deep dive into performance and scale issues in complex distributed systems, including interactions between Kubernetes and the NVIDIA software stack, to identify and resolve root causes.
Design and develop monitoring, reporting and analysis tools for performance and scale testing across software, GPU and CPU resources.
Triage, debug and root cause issues related to operating Kubernetes clusters at ultra-large scale, ensuring reliability and efficiency.
Build and maintain a high-velocity framework that enables continuous, always-on performance and scale testing via a modern CI/CD pipeline.
Document research, methodologies and results clearly and concisely, and present findings at internal and external venues, including community conferences such as KubeCon and GTC.
Engage efficiently with upstream communities—including Kubernetes, CNCF, and NVIDIA open-source projects—to validate performance and scalability of AI workloads early and help shape design and development decisions.

Requirements

8+ years of experience in Computer Architecture, Networking, Storage systems, Accelerators, and Bachelors/Masters in Engineering (preferably Electrical Engineering, Computer Engineering, or Computer Science) or equivalent experience.
Expertise in Kubernetes and familiarity with related CNCF projects.
Background in working with large scale parallel and distributed accelerator-based systems.
Expertise optimizing performance and AI workloads on large scale systems.
Experience with performance modeling and benchmarking at scale.
Proficiency in Golang/Python.
Background with the NVIDIA software ecosystem in both training and inference domains.
Expertise with at least one of public CSP infrastructure (e.g., GCP, AWS, Azure, OCI).

Ways to stand out from the crowd

Strong operational experience with any one of the Kubernetes distributions.
Prior experience scaling Kubernetes clusters to ultra-large node and object counts.
Demonstrated history of working in the open-source community.
Excellent communication and interpersonal abilities.
PhD in relevant areas.