Cluster Automation Engineer for Reproducible AI Labs

CommonAI C.I.C.; Cambridge; Hybrid; GBP 65.000 - 100.000; Vollzeit

Benefits dieser Stelle

Pension

Hybrid office near Cambridge station

Professional development opportunities

Flexible work arrangements

Zusammenfassung

CommonAI CIC is building software to configure and control real AI infrastructure for reproducible experiments in a hybrid Cambridge setting. You will enable workload placement, serving settings, resource allocation and device configuration across experimental clusters.

You will work with measurement and simulation engineers, cluster operators and AI tooling to ensure actions are validated, recoverable and reproducible, and you will leverage Linux and HPC automation expertise in a collaborative

Qualifikationen

• Experience building systems software or automation for computing infrastructure.
• Strong debugging, testing and validation skills in complex environments.
• Ability to reason about partial failure, state and permissions in running systems.
• Experience with orchestration and HPC concepts is a plus.

Aufgaben

• Identify cluster controls and expose through programmatic interfaces with clear actions.
• Automate configuration and experiment setup across the stack (workload placement, serving, resources).
• Build reliable execution with precondition checks, state verification, error handling and recovery to a known config.
• Record actions, outcomes and configurations to support reproducibility of experiments.

Kenntnisse

Strong programming and debugging

Linux

AI coding agents

CICD concepts

Partial failure reasoning

Distributed systems

Tools

Kubernetes

Slurm

NixOS/CoreOS

Terraform/OpenTofu

Jobbeschreibung

CommonAI CIC is building software to configure and control real AI infrastructure for reproducible experiments in a hybrid Cambridge setting. You will enable workload placement, serving settings, resource allocation and device configuration across experimental clusters.

You will work with measurement and simulation engineers, cluster operators and AI tooling to ensure actions are validated, recoverable and reproducible, and you will leverage Linux and HPC automation expertise in a collaborative