Senior SRE: Global GPU Infra & Production Reliability

Parasail; San Mateo (CA); Hybrid; USD 160,000 - 240,000; Full time

Job summary

Parasail is hiring Site Reliability Engineers to own and scale production infrastructure for a fast, reliable AI inference cloud. You will manage a global GPU fleet, automate operations, and improve observability to detect and recover from failures.

You will work with infra, platform, and inference engineers in a flat organization, focusing on building ambitious, reliable systems that can handle dynamic workloads at scale.

Qualifications

• Own production infrastructure with real ownership and reliability focus.
• Strong Linux fundamentals; experience with networking, storage, and containers.
• Hands-on experience running Kubernetes in production.
• Ability to write maintainable software and automation to solve infra problems.
• Systematic debugging across application, cluster, network, and hardware boundaries.

Responsibilities

• Scale a global GPU fleet by building and improving Kubernetes infrastructure across regions.
• Design better isolation, failover, and recovery to reduce customer impact.
• Automate capacity expansion, deployments, and maintenance; reduce manual work.
• Develop observability and diagnostics to reveal bottlenecks and failures.
• Own the production feedback loop: incidents, root cause, and improvements.
• Collaborate across the stack to improve performance, utilization, and security.

Skills

Kubernetes

Linux fundamentals

Distributed systems

Infrastructure operations

Automation

Debugging across layers

Tools

Kubernetes

CI/CD

Containers

Observability

Job description

Parasail is hiring Site Reliability Engineers to own and scale production infrastructure for a fast, reliable AI inference cloud. You will manage a global GPU fleet, automate operations, and improve observability to detect and recover from failures.

You will work with infra, platform, and inference engineers in a flat organization, focusing on building ambitious, reliable systems that can handle dynamic workloads at scale.