AI Fleet Platform Software Engineer

About the role

Cerebras builds the world's largest AI chip, delivering training and inference speeds over 10 times faster than GPU-based cloud services. It runs AI clusters across many data centers, and as the fleet grows, the software that monitors and manages it must scale. Cerebras is looking for a senior software engineer to build the platforms and tools that help teams understand fleet health, respond to issues and keep compute capacity available. The role spans backend services, operational tools and user-facing applications, working with Cluster Operations, infrastructure, inference, security and data center teams.

Key responsibilities

• Build and operate software for managing large fleets of AI clusters.
• Give operators clear, actionable views of cluster health, capacity, performance and ongoing issues.
• Develop services and integrations that combine data and workflows from multiple infrastructure systems.
• Automate repetitive work and improve tools for investigating incidents and restoring service.
• Design systems that stay reliable as the fleet grows and keep working through component and site failures.
• Work closely with platform users to understand their needs and make practical product and engineering decisions.
• Lead projects from design to production, measure their impact and use operational feedback to improve them.

Required profile

• 12+ years of experience building and operating production distributed systems or large-scale infrastructure software.
• Experience building control planes, fleet management systems or operational platforms.
• Experience designing APIs and data models serving multiple consumers, with attention to versioning and compatibility.
• Experience with asynchronous systems supporting idempotency, concurrency and reconciliation while avoiding unintended state drift.
• A strong testing practice, especially for software that changes the state of running production clusters.
• Sound judgment about reliability, security and observability, including behavior when dependencies are unavailable.
• Ability to take an ambiguous problem from design through production and improve it with operational feedback.

Required skills

• Strong Go; regular work with Go, Protocol Buffers and SQL (Python is an added benefit).
• Linux, containers and Kubernetes.
• Durable workflow execution, event streaming or time-series telemetry at high cardinality.
• Preferred: infrastructure software for thousands of hosts across multiple sites, incident response, hardware health or capacity management software, AWS deployment and operation, and dashboards or applications for infrastructure operators.

The position is based in the San Francisco Bay Area (Sunnyvale) or Toronto, Canada, and is full-time.