On-Prem AI Infra Engineer (Kubernetes & GPU)

Evollabs Tech; Dubai; On-site; AED 480,000 - 720,000; Full time

Job summary

Evollabs Tech seeks a Platform Engineer to own end-to-end on-prem AI infra, architecting and operating Kubernetes GPU clusters, and enabling reliable CI/CD. You will manage servers, networks, and core services, and drive observability, security, and disaster recovery.

You will build reusable IaC modules, implement GitOps, and collaborate with developers to accelerate AI workloads while maintaining strict data residency and air-gapped controls.

Qualifications

• 5+ years in DevOps, SRE, or Platform Engineering with hands-on ownership of on-prem environments.
• Advanced troubleshooting in Linux networking, storage I/O, and kernel parameters.
• Hands-on experience with physical server racks, network switches, cabling, and hardware lifecycle management.
• Experience with GPU server management, drivers, firmware, and thermal/power considerations.
• Experience designing and operating air-gapped/isolated networks with strict access controls (FreeIPA/LDAP, Kerberos, SSSD).
• Production observability setups with Prometheus and Grafana, dashboards, and alerting.
• Experience with Proxmox VE clusters, hypervisors, VM storage, and virtual networking.
• Proficiency with Docker and Kubernetes for containerized workloads and cluster ops.
• Python and Bash scripting for automation and API integrations.
• CI/CD pipelines across GitLab CI, GitHub Actions, Jenkins, etc.
• Experience integrating AI tooling and LLM endpoints (Cursor/Claude/Copilot) and cost tracking for AI infra.
• Knowledge of auditing, reproducible environments, and runbooks for engineers.

Responsibilities

• Design and operate on-prem infrastructure as code: Terraform/Ansible/Helm modules; GitOps workflows for auditable changes.
• Build and run Kubernetes for AI: multi-tenant GPU clusters, quotas, and workload isolation.
• Administer servers, networks, and core services: Linux, identity/SSO (Keycloak/LDAP), secrets (Vault), DNS/DHCP/NTP, artifact registries, mirrors.
• Enable CI/CD: collaborate with developers to create fast, reproducible pipelines and to deploy to GPU/CPU nodes.
• Observability & reliability: metrics/logs/traces with Prometheus/Grafana, define SLOs, incident response, post-mortems, and runbooks.
• Backup & disaster recovery: define RPO/RTO, backups, restores, and site failover tests.

Skills

DevOps

SRE

Platform Engineering

Linux

On-prem Hardware

GPU Hardware

Proxmox

Kubernetes

CI/CD

Terraform

Ansible

Helm

GitLab CI

NVIDIA CUDA

Vault

Keycloak/LDAP

Prometheus/Grafana

Tools

Proxmox VE

Terraform

Ansible

Helm

Kubernetes

Docker

GitLab CI

GitHub Actions

Jenkins

NVIDIA CUDA

Job description

Evollabs Tech seeks a Platform Engineer to own end-to-end on-prem AI infra, architecting and operating Kubernetes GPU clusters, and enabling reliable CI/CD. You will manage servers, networks, and core services, and drive observability, security, and disaster recovery.

You will build reusable IaC modules, implement GitOps, and collaborate with developers to accelerate AI workloads while maintaining strict data residency and air-gapped controls.