IT Search Corp
Certified AI Infrastructure and Kubernetes Platform Architect
Full Time • new york city
AI Infrastructure and Kubernetes Platform Architect – DGX Systems) Remote
Related Certifications required
6 months to 1+ yrs
$open
USC or GC req
Job Description:
We are seeking a highly skilled AI Infrastructure and Kubernetes Platform Architect with deep expertise in managing GPU-accelerated workloads on NVIDIA DGX systems. The ideal candidate will have hands-on experience with Kubernetes at the administrator, application developer, and security levels (CKA, CKAD, CKS), and will be responsible for designing, deploying, securing, and maintaining large-scale AI infrastructure powered by DGX BasePODs and SuperPODs. This role involves optimizing AI workloads, managing high-performance networking (InfiniBand), and ensuring operational excellence across NVIDIA AI systems and BlueField DPU environments.
Key Responsibilities:
Kubernetes and AI Platform Orchestration
- Architect and maintain containerized AI/ML platforms using Kubernetes on DGX systems.
- Integrate NVIDIA Base Command Manager with Kubernetes for workload scheduling and GPU resource optimization.
- Design multi-tenant GPU resource partitioning strategies using MIG (Multi-Instance GPU) to maximize hardware utilization across concurrent AI workloads.
- Implement and manage Helm charts, custom controllers, and GPU operators for scalable ML infrastructure.
DGX Infrastructure Administration
- Administer and optimize NVIDIA DGX BasePODs and SuperPODs.
- Ensure optimal GPU, CPU, and storage performance across AI clusters.
- Leverage DGX System Administration best practices for lifecycle management and updates.
- Coordinate capacity planning for DGX cluster expansion including rack power, cooling, and storage integration with NVIDIA AI Enterprise software stack.
High-Performance Networking & DPU
- Deploy, monitor, and manage InfiniBand networks using Unified Fabric Manager (UFM).
- Integrate BlueField DPUs for offloaded security, networking, and storage tasks.
- Optimize end-to-end data pipelines from storage to GPUs.
Security and Compliance
- Apply best practices from the CKS certification to harden Kubernetes clusters and AI workloads.
- Implement secure service mesh and microsegmentation with BlueField DPU integration.
- Conduct regular audits, vulnerability scanning, and security policy enforcement.
Automation & Monitoring
- Automate deployment pipelines and infrastructure provisioning with IaC tools (Terraform, Ansible).
- Monitor performance metrics using GPU telemetry, Prometheus/Grafana, and NVIDIA DCGM.
- Troubleshoot and resolve complex system issues across hardware and software layers.
- Implement MLOps workflows integrating KubeFlow Pipelines, NVIDIA Triton Inference Server, and model registry tooling to support end-to-end model training and production deployment.
Required Skills and Qualifications:
- CKA, CKAD, CKS certifications – demonstrating full-stack Kubernetes expertise.
- Proven experience with NVIDIA DGX systems and AI workload orchestration.
- Hands-on expertise in InfiniBand networking, UFM, and BlueField DPU administration.
- Strong scripting and automation skills in Python, Bash, YAML.
- Familiarity with Base Command Manager, NVIDIA GPU Operator, and KubeFlow is a plus.
- Ability to work across teams to support ML researchers, DevOps engineers, and infrastructure teams.
==================================================
Compensation: $110.00 - $150.00 per hour
(if you already have a resume on Indeed)