Senior GPU Infra Reliability Engineer - Remote

Luma AI; United States; Remote; USD 180,000 - 240,000; Full time

Job summary

Luma AI is seeking a Senior SRE to own the GPU infrastructure used for training and inference across on-prem, AWS, and OCI. You’ll be hands-on, close-to-the-metal, tackling GPU, networking, and kernel-level failures and driving reliability at scale.

You’ll tune Linux performance, build automation in Python/Go/Bash, and participate in security and compliance efforts while partnering with NVIDIA and other vendors in a fast, unstructured environment.

Qualifications

• 5+ years as SRE, production, or infrastructure engineer in a fast-paced, large-scale environment.
• Deep Linux expertise, containerized systems, and kernel/OS-level performance debugging.
• Hands-on experience with cloud platforms (AWS or OCI) and high-speed networking.

Responsibilities

• Take end-to-end ownership of production GPU clusters for training and inference across AWS/OCI.
• Join re-architecture sessions to redesign systems for higher efficiency and scale.
• Tune Linux performance deeply at the OS and kernel level.
• Build automation in Python, Go, or Bash to manage, monitor, and self-heal infrastructure.
• Serve as final escalation for GPU, networking (InfiniBand/RDMA), and system failures with vendors like NVIDIA.
• Support security certifications (SOC 2 Type 1/2, ISO) with strong infrastructure practices.

Skills

Linux expertise

Containerized systems

Performance debugging

Networking experience

Tools

Terraform

Airflow

Ray

AWS

OCI

Kubernetes

Job description

Luma AI is seeking a Senior SRE to own the GPU infrastructure used for training and inference across on-prem, AWS, and OCI. You’ll be hands-on, close-to-the-metal, tackling GPU, networking, and kernel-level failures and driving reliability at scale.

You’ll tune Linux performance, build automation in Python/Go/Bash, and participate in security and compliance efforts while partnering with NVIDIA and other vendors in a fast, unstructured environment.