AI Infra Engineer – HPC Clusters & Slurm/K8s

Veeda AI; Seattle (WA); On-site; USD 180,000 - 240,000; Full time

Job summary

Veeda AI in Seattle is seeking an infrastructure engineer to design, deploy, and operate state-of-the-art GPU clusters and data services for multimodal AI workloads. You will manage compute health, interconnects, storage, and scheduler resources across Slurm and Kubernetes environments.

The role emphasizes security, observability, and scalability, collaborating with model and data teams to plan capacity and tooling. A strong Linux/HPC background and IaC automation are essential for success.

Qualifications

• Bachelor’s degree in Computer Science, Computer Engineering, or equivalent hands-on experience in HPC infra.

Responsibilities

• Design, deploy, and operate Slurm-on-Kubernetes GPU clusters with hardware health monitoring.
• Implement access control for source code, clusters, data storages, and dev tools.
• Build distributed data transmission and caching services for timely data delivery.
• Develop telemetry pipelines with Prometheus, Grafana, DCGM, and Redfish.
• Forecast resources with Model and Data teams and run PoCs with vendors.
• Prepare infra-OPEx forecasts and provide guidance on spend decisions.
• Lead end-to-end projects from specs to CI/CD and on-call production support.

Skills

Linux networking

Kernel tuning

Automation tooling

Python scripting

HPC clusters

Slurm Kubernetes admin

PyTorch NCCL

Cloud workload profiling

Security IaC

Education

Bachelor's degree

Tools

Ansible

Terraform

Helm

GitLab CI

Docker

Prometheus

Grafana

OpenTelemetry

Slurm

Kubernetes

Ceph

Lustre

NFS

NVMe-oF

Job description

Veeda AI in Seattle is seeking an infrastructure engineer to design, deploy, and operate state-of-the-art GPU clusters and data services for multimodal AI workloads. You will manage compute health, interconnects, storage, and scheduler resources across Slurm and Kubernetes environments.

The role emphasizes security, observability, and scalability, collaborating with model and data teams to plan capacity and tooling. A strong Linux/HPC background and IaC automation are essential for success.