Full job description

Securing NVIDIA's EDA clusters is not a matter of deploying a standard security stack. The workloads are too large, too long-running, and too performance-sensitive for an off-the-shelf playbook to simply apply. That constraint is the job. You'll be an individual contributor on Infrastructure Security Engineering for EDA Clusters, the team that implements security controls across the compute, storage, and networking that NVIDIA's hardware engineers design silicon on - spanning on-prem data centers and multi-cloud. You will sit alongside the engineers who run these clusters 24x7, close enough to deploy a control, investigate a behavior, and change the running environment yourself.

This is hands-on build work, not control review. You'll decide where conventional tooling applies, where it actively cannot, and then engineer the alternative. If your depth is in large-scale compute, networking, or storage infrastructure at a hyperscaler, a security product company, or anywhere fleets and fabrics are the day job, and you've carried real accountability for the security properties of those systems, you're who we want to talk to. We don't expect fluency in every security product category. We expect you to own outcomes and pick up the remaining domain breadth fast. Security engineering here is engineering: code, instrumentation, and secure-by-default infrastructure - not tickets, exceptions, and spreadsheets.

What You'll Be Doing:

Build and operate security controls and telemetry for EDA compute, storage, and networking - design, ship to production, then own them.
Engineer around hard constraints. Protect multi-day, high-value compute jobs without degrading them. Where heavyweight in-node agents are not viable, build the workable answer instead.
Make the network and platform the control plane. Protection can't depend solely on what runs inside the node.
Instrument for detection. Collect and normalize security signals from the fleet into our SIEM, and build the detections and dashboards that turn it into something actionable. Log collection alone is not monitoring.
Investigate and respond. Chase down EDA-specific threats and exposure, and execute containment and remediation hands-on with infrastructure and service owners.
Automate enforcement and evidence. Infrastructure-as-code and policy-as-code so secure configuration is the default state, with compliance evidence produced as an engineering output rather than a manual exercise.
Run the lifecycle of the security technologies you own: selection, rollout, operation, and eventual replacement.
Work across boundaries. Partner directly with data center operations, IT storage and networking, network engineering, CSP teams, and NVIDIA's central security organization on shared outcomes.