We are looking for a DevOps Engineer to own and scale the infrastructure behind our AI systems.
What you'll do
Own cluster reliability, availability, monitoring, alerting, capacity planning, and incident response.
Configure, maintain, and troubleshoot GPU servers, networking equipment, switches, InfiniBand fabrics, storage, and related infrastructure.
Deploy, maintain, and improve infrastructure around Slurm, Kubernetes, and GPU workload orchestration.
Manage and improve our cloud environment, including compute, networking, storage, IAM, security groups, cost visibility, and production infrastructure.
Build, maintain, and improve CI/CD pipelines.
Work closely with engineers to understand infrastructure bottlenecks and improve development, training, inference, and deployment velocity.
Contribute to security best practices across cloud, network, CI/CD, access control, secrets management, and production systems.
Requirements: Strong hands-on experience managing Linux-based infrastructure in production environments.
Experience with classical DevOps responsibilities: CI/CD, cloud infrastructure, monitoring, automation, deployment systems, and production operations.
Familiarity with Kubernetes in real-world production environments.
Experience operating GPU clusters, HPC environments, AI infrastructure, or other high-performance compute systems.
Familiarity with Slurm or similar workload schedulers.
Practical knowledge of networking, including switch configuration, routing, VLANs, firewalls, high-throughput networking, and debugging network issues.
Experience with InfiniBand, RDMA, high-performance interconnects, or similar low-latency networking technologies.
Ability to troubleshoot across the full stack: hardware, firmware, BIOS, Linux kernel, drivers, CUDA, containers, networking, storage, cloud infrastructure, CI/CD, and schedulers.
Strong scripting and automation skills using Bash, Python, Ansible, Terraform, or similar tools.
A strong ownership mindset: you notice problems, fix root causes, automate repetitive work, and leave systems more reliable than you found them.
This position is open to all candidates.