To ensure your application receives a thorough and fair review, you must complete the application form below: https://forms.gle/VZTwn2JQ49pZTQXU8

This is a joint hire by Strontium and IntellifAI Labs. Strontium and IntellifAI Labs are collaborating to bring state-of-the-art AI technologies to everyday users.

About us

Strontium is one of the fastest growing AI companies in Canada: fully bootstrapped, 15+ years in the game, serving 100,000+ customers across 60+ countries. We build and sell AI-powered creative tools directly to consumers. For example, here are two of our products:

- VideoExpress.ai: our video generation product, rated 4.8★ on Capterra with 6000+ reviews and #1 in the category: https://www.capterra.com/p/10019949/VideoExpress/reviews/
- Artistly.ai: our image generation product, rated 4.8★ on TrustPilot with 1000+ reviews: https://ca.trustpilot.com/review/artistly.ai

We run our own image, video and audio models on our own GPU fleet. We ship rapid, iterate constantly, and compete with companies 100x our size.

The role

Our models run on a large fleet of GPU servers that we operate ourselves, both on-premises and in the cloud. We need one senior server admin who keeps that fleet secure, healthy and predictable. You start with day-to-day operations and steadily take over fleet management, deployments and hardening.

You will use agentic tools (Claude Code, Codex) every day, and we pay for them. But on a production fleet, a command you don't understand is a risk no matter who wrote it. We want the person who reads the command before running it, checks the live server before trusting the document, and has a rollback ready before making the change.





Full-time · Remote · Flexible hours (reasonable overlap with EST) · Competitive compensation

What you'll do

- Run the fleet. Linux on bare-metal and cloud GPU servers: provisioning, patching, NVIDIA drivers, disks, networking, monitoring and alerting. Configuration with Ansible; services under Docker Compose, Supervisor and systemd.
- Make deployments and scaling safe. Containers, CI/CD with GitHub Actions, staged rollouts, health checks and rollbacks. Kubernetes where it earns its place.
- Harden everything. Secrets management, SSO and 2FA, VPN, least-privilege access, key rotation and audit trails. You close a finding and then prove it is closed.
- Protect the data. Postgres, Redis and S3-compatible object storage, with backups you have actually restored.

Minimal qualifications

- Hands-on Linux server administration in production, including bare metal. You have been the person responsible when a server went down.
- NVIDIA GPU servers: you have installed and fixed NVIDIA drivers, CUDA and the container toolkit on Linux.
- Containers in depth: Docker, Compose and Kubernetes. You can explain what happens on the host when a container starts, and what a volume is before you delete one.
- Configuration management (Ansible or similar)



across dozens of hosts or more.
- Python and Bash good enough to read, debug and fix the scripts that run a fleet.
- Networking and security fundamentals: firewalls, VPN, TLS, SSH, SSO, secrets.
- Monitoring and alerting with Prometheus, Grafana or similar.
- Agentic coding tools used with judgment. You can show a case where the AI was wrong and how you caught it.
- Agency on reversible things, caution on irreversible ones, and you know which is which. You find the second path and get expert help without waiting for permission, and you ask before a step you cannot undo.
- Clear written English. Your updates say what is done, what is left, what is blocked and what is next.

Nice to have

- GPU fleet operations at scale: thermals, failed-card triage, GPU health monitoring across many servers.
- Google Cloud or AWS experience on top of bare metal: IAM, secrets management, GPU instances.
- Postgres administration: replication, point-in-time recovery.
- MinIO or other S3-compatible storage at scale.

This role is not for you if

- Your experience is only managed cloud consoles and certifications.
- You run commands from an AI tool or a document in production without being able to say what each part does.
- You fix things without writing down what you changed.
- You hand a blocker back instead of finding a second path.

How to apply

Complete the application form linked at the top. In the form, begin your answer to the question about the largest fleet you ran with the word "bare-metal" so we know you read this post.

For questions, contact [email protected].

📌 Senior DevOps Engineer / Server Admin (GPU Infrastructure) (India)
🏢 IntellifAI Labs
📍 India