Infrastructure Engineer (GPU & Compute)
Owns GPU diagnostics, validation workflows, and automation for bare-metal infrastructure supporting AI/ML workloads. Requires 5+ years in systems engineering with strong Linux, Python, and NVIDIA tools expertise.
About the job
What You'll Do
Systems, Image & Validation Infrastructure
- Own and evolve systems for image management, deployment, and validation across bare-metal infrastructure
- Run and maintain test clusters used for system validation, diagnostics, and bring-up
- Validate firmware, drivers, and OS images across compute and GPU-enabled systems
- Support hardware qualification efforts for next-generation platforms
GPU Diagnostics & Performance
- Own GPU diagnostics and validation workflows across large-scale infrastructure
- Diagnose and resolve complex issues across GPUs, drivers, OS, and hardware layers
- Analyze system and GPU performance using tools such as NVIDIA DCGM
- Identify failure patterns and drive improvements in system stability and validation coverage
Automation & Tooling
- Build and maintain automation for provisioning, validation, and system bring-up
- Develop Python-based tools and workflows to improve efficiency and reduce manual operational overhead
- Improve the reliability, repeatability, and scalability of image pipelines and validation systems
Systems & Operations
- Manage and operate Linux-based systems in production and validation environments
- Manage virtualization technology
- Support bare-metal provisioning workflows, including PXE and image-based systems
- Interface with hardware management systems (e.g., IPMI, Redfish) for monitoring and debugging
Cross-Functional Collaboration
- Partner with Infrastructure, Hardware, and Data Center teams on system bring-up and validation
- Collaborate with platform and ML teams to ensure systems meet workload requirements
- Contribute to best practices for provisioning, diagnostics, and lifecycle management of infrastructure
What You'll Need
Required Qualifications
- 5+ years of experience in infrastructure engineering, systems engineering, or related roles
- Strong Linux systems experience in production environments
- Hands-on experience with GPU-enabled systems and tools such as NVIDIA DCGM
- Familiarity with bare-metal provisioning and system bring-up workflows
- Proficiency in Python or similar scripting/programming languages for automation
- Ability to debug complex issues across hardware, OS, GPUs, and system software
Ideal Experience
- Experience with high-performance interconnects (e.g., InfiniBand, NVLink)
- Experience with PXE boot environments, LiveCD systems, or image-based provisioning workflows
- Experience with hardware management interfaces such as iDRAC, IPMI, or Redfish
- Data center operations experience, including working with physical hardware
- Experience supporting AI/ML or HPC workloads at scale
- Experience with GPU validation frameworks or large-scale hardware qualification processes
Skills
Linux, Python, Nvidia Dcgm, Bare-Metal Provisioning, Pxe, Ipmi, Redfish, InfiniBand, Nvlink, Virtualization
Similar jobs
DevOps / SRE jobs
Benchling
San Francisco, CABuild developer-experience tooling and release systems within Benchling’s Platform team, helping engineering teams develop, test, package, and ship high-quality software rapidly. The role requires 4+ years of software engineering experience, web framework expertise, strong problem-solving, and effective cross-functional communication.
Baseten
San Francisco, CA
Leads global GPU capacity management across acquisition, orchestration, infrastructure automation, and incident response. The role requires 5+ years of experience, deep Kubernetes expertise, production Go or Python skills, and the ability to balance reliability with unit economics.
Mercor
San Francisco, CA
Own Mercor’s internal identity and cloud platform infrastructure as code, automating provisioning, access management, secrets, and employee lifecycle workflows. The role requires production Terraform, Okta, SCIM, and multi-cloud IAM experience, plus strong automation, incident response, and documentation skills.
Ramp
New York, NYProduction Engineer responsible for building and operating scalable infrastructure, driving reliability and architecture initiatives, and enabling product teams through platform tooling. Requires software engineering experience, distributed-systems expertise, cloud experience, and cross-team technical leadership.
Roboflow
New York, NYInfrastructure Engineer responsible for securing, scaling, and operating cloud infrastructure and machine-learning platforms across a distributed startup. The role requires Kubernetes, infrastructure-as-code, cloud operations, CI/CD, programming, observability, and security experience.