Staff Software Engineer - Reliability

About the position

The SRE role sits with the Infrastructure and Research teams and is responsible for our GPU clusters. We run on '1000s of H100 GPUs across multiple providers and clusters for Training, Data Processing, and Inference. We need a highly skilled SRE to ensure those clusters are healthy and to build the monitoring and management tools we need to make full use of them. Successful candidates will want to get extremely in the weeds solving performance and maintenance problems in our clusters.

Responsibilities

• Collaborate with researchers and engineers to specify the availability, performance, correctness, and efficiency requirements of the current and future versions of our GPU infrastructure.
• Work with multiple GPU cloud providers to scale up, scale down, maintain and monitor our 1000's GPUs in many clusters.
• Design and implement solutions to ensure the scalability of our infrastructure to meet rapidly increasing demands.
• Implement and manage monitoring systems to proactively identify issues and anomalies in our production environment.
• Implement fault-tolerant and resilient design patterns to minimize service disruptions.
• Build and maintain automation tools to streamline repetitive tasks and improve system reliability.
• Participate in an on-call rotation to respond to critical incidents and ensure 24/7 system availability alongside other infrastructure developers.
• Develop and maintain service level objectives (SLOs) and service level indicators (SLIs) to measure and ensure system reliability.

Requirements

• Proven work experience 10+ yrs as a reliability engineer, production engineer, infrastructure software engineer or a similar role in a fast-paced, rapidly scaling company.
• Strong proficiency in GPU cloud infrastructure, including the underlying concepts of scheduling, scaling, cloud storage, networking and security.
• Proficiency in programming/scripting languages.
• Experience with containerization technologies and container orchestration platforms like Kubernetes or equivalent.
• Knowledge of IaC tools such as Terraform or CloudFormation or equivalent.
• Excellent problem-solving and troubleshooting skills.
• Strong communication and collaboration skills.
• Experience with observability tools; examples include DataDog, Prometheus, Grafana, Splunk and ELK stack or similar.
• Knowledge of security best practices in cloud environments.

Nice-to-haves

• Good to have experience as an SRE within the AI/ML space is strongly preferred.

Benefits

• Comprehensive benefits package, all subject to eligibility.