HPC & AI Platform Administrator

Lilly; South San Francisco (CA); Hybrid; USD 141,000 - 231,000; Full time

Benefits offered by this job

Bonus program

Comprehensive benefits

Job summary

Lilly in Silicon Valley is seeking a seasoned HPC Systems Administrator to build and operate scalable AI and high-performance computing platforms powering advanced machine learning research.

You will partner with AI scientists, engineers, and infrastructure teams to enable training, inference, and experimentation across GPU, cloud, and on‑prem environments, while advancing automation and reliability in a hybrid, three-days onsite/two-days remote model.

Qualifications

• Bachelor's degree in CS/EE/Systems Engineering or related field.
• 5 years deploying, administering, or supporting large-scale HPC or distributed computing environments.
• Experience with GPU/AI infrastructure in cloud and on-prem environments.
• Knowledge of GPU hardware lifecycle, monitoring, and observability.

Responsibilities

• Build and operate scalable AI and HPC platforms powering ML workloads.
• Partner with AI scientists, engineers, and domain experts to enable training, inference, and experimentation.
• Improve productivity, performance, and access to computing resources while ensuring reliability and availability.
• Drive automation, standardization, and cost-efficient operations.

Skills

Linux administration

Python scripting

Bash scripting

Ansible

Slurm/Grid Engine

Kubernetes

Docker

Cloud platforms

Automation

Collaboration

Education

Bachelor's in CS/EE/SE or related field

Tools

Slurm

Grid Engine

Ansible

Kubernetes

Docker

AWS

Azure

GCP

Job description

Lilly in Silicon Valley is seeking a seasoned HPC Systems Administrator to build and operate scalable AI and high-performance computing platforms powering advanced machine learning research.

You will partner with AI scientists, engineers, and infrastructure teams to enable training, inference, and experimentation across GPU, cloud, and on‑prem environments, while advancing automation and reliability in a hybrid, three-days onsite/two-days remote model.