Site Reliability Engineer
Gruve
About Gruve
Gruve is an innovative software services startup dedicated to transforming enterprises to AI powerhouses. We specialize in cybersecurity, customer experience, cloud infrastructure, and advanced technologies such as Large Language Models (LLMs). Our mission is to assist our customers in their business strategies utilizing their data to make more intelligent decisions. As a well-funded early-stage startup, Gruve offers a dynamic environment with strong customer and partner networks.
Position Summary:
This role owns reliability, automation, and stability across GPU-backed inference systems, Kubernetes environments, and ML Ops stacks. You will work with other Site Reliability engineers to set up, maintain, and troubleshoot the stack from bare metal through the application layer.
Key Responsibilities:
- Operate Kubernetes clusters across Rafay, OpenShift, Mirantis, native K8s, and Rancher.
- Manage GPU computing stack: CUDA, ROCm, DGX/HGX, Nemotron, NIMs.
- Maintain ML Ops: Kubeflow, model upgrades, and lifecycle management.
- Implement Infrastructure-as-Code and automation for platform repeatability and resilience.
- Support SLO management, incident response, and capacity planning.
Basic Qualifications:
- 3–5 years of SRE or DevOps experience.
- Strong Kubernetes operations and Linux skills.
- Proficiency with IaC (Terraform, Pulumi, or Ansible).
- Strong GPU or ML Ops skills.
Preferred Qualifications:
- Working experience with job schedulers (Slurm, kserve).
- Working experience with CNI networking (Calico, SR-IOV).
Why Gruve
At Gruve, we foster a culture of innovation, collaboration, and continuous learning. We are committed to building a diverse and inclusive workplace where everyone can thrive and contribute their best work. If you’re passionate about technology and eager to make an impact, we’d love to hear from you.
Gruve is an equal opportunity employer. We welcome applicants from all backgrounds and thank all who apply; however, only those selected for an interview will be contacted.