AI Infrastructure & Managed Service Director

Role Overview

Lead the end-to-end managed services and infrastructure operations for NVIDIA GB300-powered AI infrastructure, ensuring high availability, performance, security, scalability and operational efficiency of the AI factory.

The role will own the operational lifecycle across GPU compute, networking, storage, Kubernetes/container platforms, data center infrastructure, observability, automation and AI workloads, working closely with NVIDIA, hyperscalers, technology partners and enterprise customers.

Key Responsibilities

1. GPU Infrastructure Operations

* Own day-to-day operations of NVIDIA GB300 GPU infrastructure and AI compute platforms.

* Manage GPU clusters, including NVL72-based systems, GPU nodes, high-speed networking and associated infrastructure.

* Ensure GPU availability, utilization, performance and capacity meet business SLAs.

* Establish GPU health monitoring, predictive maintenance and proactive fault management.

* Manage firmware, drivers, CUDA stack and infrastructure lifecycle in coordination with NVIDIA best practices.

2. AI Factory Infrastructure Management

* Manage the complete AI infrastructure stack across:

  * GPU compute

  * High-performance networking

  * Storage

  * Kubernetes/container platforms

  * Virtualization/bare-metal environments

  * Data center facilities and power/cooling dependencies

* Establish operational standards for AI Factory environments.

* Drive infrastructure capacity planning for rapidly increasing AI workloads.

3. Managed Services

* Define and operate 24×7 managed services for GPU infrastructure.

* Establish L1/L2/L3 support, NOC/SOC integration, incident, problem and change management.

* Define SLAs, OLAs, KPIs and service governance.

* Lead major incident management and service restoration.

* Establish operational runbooks and standardized troubleshooting procedures.

4. Observability & AIOps

* Build a unified observability model covering GPU → server → network → storage → Kubernetes → application/workload.

* Monitor GPU utilization, temperature, memory, power, ECC errors, NVLink/NVSwitch health and workload performance.

* Implement AI-driven anomaly detection and predictive operations.

* Integrate platforms such as Datadog, ServiceNow and Grafana for unified operations and executive reporting.

5. Automation & Autonomous Operations

* Automate provisioning, patching, health checks, remediation and capacity management.

* Develop self-healing capabilities for common infrastructure failures.

* Use AI agents/AIOps to move from reactive incident management toward predictive and autonomous operations.

* Establish Infrastructure-as-Code and automated configuration management.

6. GPU Capacity & Performance Management

* Monitor GPU utilization and workload efficiency.

* Optimize GPU allocation and scheduling.

* Identify stranded/underutilized GPU capacity.

* Establish FinOps / GPU-as-a-Service billing and chargeback/showback models.

* Provide capacity forecasts based on customer workload demand.

7. Security & Compliance

* Ensure security of GPU clusters, management planes, containers and AI workloads.

* Implement vulnerability, patch and configuration management.

* Work with cybersecurity teams on zero-trust, privileged access and workload security.

* Ensure compliance with enterprise and data-center security standards.

8. Vendor & Customer Management

* Act as the operational interface between the customer, NVIDIA and technology partners.

* Manage OEM, networking, storage, cloud and software vendors.

* Drive technical governance and service reviews with customer CIO/CTO organizations.

* Own operational escalations and service improvement programs.

Required Technical Expertise

* NVIDIA GB300 / Blackwell architecture

* NVIDIA NVL72 / GPU clusters

* CUDA ecosystem

* NVIDIA networking / NVLink / NVSwitch

* InfiniBand / high-performance Ethernet

* Kubernetes and containers

* Linux

* High-performance computing

* AI/ML infrastructure

* GPU monitoring and telemetry

* Data center infrastructure

* Storage architecture

* Automation / Infrastructure-as-Code

* AIOps and observability

* ServiceNow / ITSM

* Datadog / Grafana or equivalent observability platforms