Senior Software Engineer, Cloud-Native Stack – CSP Engagements

NVIDIA•Austin, CA

14d

About The Position

We are developing advanced multi-rack, multi-tenant AI/ML datacenters with NVIDIA GB200, and upcoming GB300 GPUs. NVIDIA seeks a Senior Software Engineer for our CSP (Cloud Service Provider) Engagements team to focus on the cloud-native stack for datacenter products like GB200. In this role, You will define customer workflows, prototype stack enhancements, and debug the toughest Kubernetes + Slurm issues in multi-rack, multi-tenant AI datacenters. You'll tackle complex scheduling challenges across racks, tenants, and clouds as part of the CSP engagements team.

Requirements

Strong source-level expertise in Kubernetes internals (scheduler, CRI/CNI/CSI, operators) and Slurm (federation, power-save, plugins).
Hands-on experience integrating next-gen GPUs (Blackwell/GB200/GB300) or comparable accelerators into containerized clusters.
Proven track record debugging large-scale, cloud-native stacks across networking (RDMA/RoCE), storage, and control planes.
Customer-facing engineering or solutions-architect background: requirements gathering, PoC ownership, roadmap influence.
Familiarity with CI/CD (GitHub Actions, Tekton), observability (Prometheus, OpenTelemetry), and infrastructure-as-code.
Excellent communication-able to switch between deep technical detail and high-level business impact.
6+ years of professional software development experience in distributed systems (Go, Rust, C/C++ or Python for tooling).
BS or MS (or equivalent experience) in Computer Engineering, Computer Science, or related field.

Nice To Haves

Upstream contributions to Kubernetes, Slurm, Volcano, or similar projects.
Experience with GPU computing (CUDA), deep learning workloads

Responsibilities

Perform deep-dive debugging of multi-rack, multi-tenant clusters: scheduler behavior, container runtime issues, device-plugin crashes, RDMA/IB fabric anomalies, etc.
Gather customer requirements and prototype feature extensions for Kubernetes operators, Slurm plugins, and custom micro-services that expose new GPU capabilities.
Drive joint architecture reviews and “whiteboard” sessions with CSP and internal platform teams; convert findings into RFCs and upstream pull requests.
Create reproducible testbeds (Helm/Ansible/Terraform) that mirror customer environments; automate validation and benchmark suites.
Deliver technical collateral-design docs, how-to guides, demo scripts-and present at customer on-sites, KubeCon, and SlurmUG.
Collaborate with AE, FAE, and Solution Architect teams to deliver integrated customer solutions and technical documentation.