This role involves designing, implementing, and optimizing GPU-accelerated container platforms for high-performance workloads such as AI/ML and HPC across hybrid or on-prem environments. The position requires deep expertise in both NVIDIA and Kubernetes ecosystems.
Responsibilities:
Architect and operate Kubernetes clusters optimized for GPU workloads.
Develop and maintain custom Kubernetes operators and controllers.
Integrate NVIDIA device plugins and optimize GPU utilization.
Collaborate with HPC, ML, and DevOps teams for high-throughput cluster performance.
Drive observability and telemetry integrations.
Implement secure multi-user and multi-namespace GPU isolation.
Maintain CI/CD pipelines for Kubernetes infrastructure.
Contribute to infrastructure-as-code using tools like Terraform and Helm.
Participate in performance tuning and incident response.
Requirements:
Extensive experience with Kubernetes in production environments.
Proficiency in Go or Python for operator development.
Deep understanding of Kubernetes internals and GPU-intensive workloads.
Hands-on experience with Helm, Kustomize, and GitOps workflows.
Familiarity with CNI plugins, especially NVIDIA CNI and Multus.
Experience with monitoring GPU metrics using Prometheus and DCGM Exporter.
26-00812
Numbers & Facts
Location
Dallas, TX
Salary
$100–$120 Per Hour
Skills
Artificial Intelligence (AI)unmatched
Continuous Deployment/Deliveryunmatched
Continuous Integrationunmatched
DevOpsunmatched
GPU (Graphics Processing Unit)unmatched
High Throughputunmatched
Incident Responseunmatched
Metricsunmatched
Performance Tuning/Optimizationunmatched
Production Systemsunmatched
Python Programming/Scripting Languageunmatched
Telemetryunmatched
Web Client Plug-insunmatched
🎯
Be found by employers
5,500+ employers search our resume database daily. Add yours to get found by recruiters looking for candidates like you.
Level up your application
Professional resume templates
Browse dozens of recruiter approved resume templates, layouts and formats. Choose your favorite and make it your own in minutes.