Senior Cloud Infrastructure Engineer - T-Labs

About Trener
Trener is building the software foundation for a new generation of industrial robotics. We combine advanced AI, intuitive programming, and pre-trained skill models to make robots easier to deploy, operate, and adapt to real-world industrial work. With teams in San Jose and Trondheim, we build software that connects cloud infrastructure, developer platforms, and deployed robotic systems.
About the Role
T-Labs is Trener's research organization. Our systems are on robot learning, vision, and motion planning, and they need GPU compute, training infrastructure, and large dataset management to do it. That infrastructure runs on AWS. The foundations are in place: networking, GPU development machines, dataset storage and replication, container registries, and managed training. We now want someone who owns it full time.
You will own the T-Labs AWS environment. You will work as a peer to a cloud engineer on our product engineering team, who owns our GCP environment. The two of you agree on shared conventions and review each other's infrastructure changes, and maintain good foundations for our cloud infrastructure. You will also work closely with our delivery infrastructure engineer, and the greater engineering team.
This is a senior individual contributor role reporting to the CTO. You will not have direct reports at the start. As the team grows, this role may take on technical leadership of cloud infrastructure.
What You Will Own
- The T-Labs AWS environment: account structure, VPCs, IAM, and the OpenTofu modules and Atmos configuration that describe them.
- Compute for research: GPU development machines and training infrastructure, sized for what researchers need and shut down when they don't.
- Data infrastructure: storage for datasets and model artifacts, replication, backup, and lifecycle rules as data volume grows.
- Container images for training: registries and replication for the images research workloads run on.
- Infrastructure delivery: reviewed plan/apply in CI with drift detection, built together with your GCP counterpart so both clouds work the same way.
- Cost visibility: tagging, budgets, and alerts that make GPU and storage spend visible to the people creating it.
- Self-service for researchers: documented, repeatable ways for ML engineers to get compute and storage without filing a request.
Requirements
- Production AWS operations. You have operated production AWS infrastructure as a cloud, platform, infrastructure, or site reliability engineer.
- Terraform or OpenTofu at module-authoring depth. You have written and versioned reusable modules, managed state across multiple environments, and handled the lifecycle of real infrastructure over time.
- AWS networking and IAM depth. VPC design, security groups, IAM roles and policies, cross-account access, and debugging access and connectivity problems.
- Infrastructure for ML or data-heavy workloads. You have run GPU compute, training jobs, or large-scale data storage for engineers or researchers, and you know where the cost goes.
- Containers. Building and running container images for compute workloads.
- Python and Bash for automation, diagnostics, and glue, plus Linux fluency.
- Clear written communication in English. Infrastructure decisions here are documented and reviewed in writing, often across time zones.
Experience That Will Help You Succeed
- Infrastructure composition tooling such as Atmos, Terragrunt, or Terraspace.
- Managed training platforms such as SageMaker or Vertex AI, or GPU scheduling with Kubernetes, Ray, or Slurm.
- GPU cost control: spot capacity, idle shutdown, right-sizing.
- Building CI for infrastructure: automated plan/apply, policy checks, drift detection.
- Production experience on GCP.
- Working closely with ML researchers or robotics engineers as your main users.
What We Offer
The opportunity to build the infrastructure behind robot learning research, as a founding member of a cloud team. A senior technical role with direct influence over how our research organization works, reporting to the CTO.
A growing international engineering organization with teams in the United States and Norway.

Employment type
FullTime

Work arrangement
No