Senior SRE: Global GPU Inference Infra & Reliability

Kindredventures; San Mateo (CA); On-site; USD 150,000 - 240,000; Full time

Job summary

You’ll own production reliability end-to-end, from capacity planning to incident response, with a focus on scalable, observable, and secure infrastructure as inference workloads grow.

Qualifications

• Experience building and operating production infrastructure or distributed systems with real ownership of reliability.
• Strong Linux fundamentals and practical knowledge of networking, storage, and containers.
• Hands-on experience running Kubernetes in production.
• The ability to write maintainable software and automation to solve infrastructure problems.
• A systematic approach to debugging problems that cross application, cluster, network, and hardware boundaries.
• Good judgment about when to move quickly, when to simplify, and where reliability matters most.
• The initiative to take a problem from investigation through implementation and work closely with teammates along the way.

Responsibilities

• Scale a global GPU fleet by building and improving the Kubernetes infrastructure behind provisioning, networking, storage, and service deployment across providers and regions.
• Design better isolation, failover, and recovery so hardware and infrastructure failures have less impact on customers.
• Automate capacity expansion, deployments, and maintenance, eliminating manual work and making changes safer.
• Develop observability and diagnostics that reveal bottlenecks, surface failures, and help engineers act quickly.
• Respond to incidents, diagnose root causes, and translate learnings into stronger systems.
• Collaborate across the stack to improve performance, utilization, security, and reliability as inference demand grows.

Skills

Production infra ownership

Linux fundamentals

Networking & storage

Kubernetes in prod

Automation & scripting

System observability

Tools

Kubernetes

CI/CD tools

Infrastructure as Code

Monitoring tools

Job description

You’ll own production reliability end-to-end, from capacity planning to incident response, with a focus on scalable, observable, and secure infrastructure as inference workloads grow.