Cloud AI architect
Summary
Hands-on private cloud AI architect in Abu Dhabi (on-site) who designs, deploys, and operates bare-metal on-prem AI infrastructure: Red Hat OpenShift/Kubernetes on NVIDIA SXM GPUs over InfiniBand, serving and fine-tuning open-weights LLMs with vLLM, Triton, and NIM, plus Ansible/Terraform automation and Prometheus/Grafana monitoring.
Hiring: Cloud AI Architect (Hands-On / Bare-Metal Infrastructure)
Location: Abu Dhabi, UAE (On-Site)
Employment Type: Full-Time, Permanent
About the Opportunity
We are seeking a deeply technical, hands-on Private Cloud AI Architect to support a large-scale enterprise operating in a high-security, mission-critical sector.
In this role, you will own the architecture, bare-metal deployment, and day-to-day stability of a high-density, sovereign on-premises AI infrastructure environment.
You will bridge physical bare-metal systems engineering with open-source container orchestration building a high-throughput internal AI utility engineered to host, fine-tune, and serve massive open-weights models (such as Llama, Mistral, and DeepSeek) entirely within secure, physical data center facilities.
Key Responsibilities
- Deploy Red Hat OpenShift via CLI using CoreOS Ignition. Configure multi-tenant RBAC, NMState (100G/400G bonding, VLANs), and Kubeflow/Tekton pipelines.
- Deploy and tune NVIDIA NIM, vLLM, Triton, and TensorRT-LLM on next-gen SXM GPUs. Configure MIG and fractional GPU slicing via CLI.
- Tune Linux kernels (sysctl, AMD EPYC NUMA balancing), patch NVIDIA drivers/NCCL, and validate GPUDirect RDMA on Quantum InfiniBand fabrics.
- Write Ansible playbooks and Terraform for firmware updates and cluster scaling. Manage local S3 connectors and high-speed NVMe (xfs/ext4) scratch storage.
- Build Prometheus/Grafana dashboards with NVIDIA DCGM to monitor InfiniBand links, memory, and rack-level power/thermal limits.
Requirements
- 8–10+ years in enterprise infrastructure, with 3–5+ years hands-on building bare-metal Kubernetes/OpenShift or on-prem HPC/AI clusters.
- Practical experience deploying NVIDIA NIM/NGC and tuning vLLM, Triton, DeepSpeed, or Ray.
- Deep CLI mastery (oc, kubectl) of bare-metal OpenShift and expert RHEL/Ubuntu administration.
- Hands-on experience with NVIDIA Quantum InfiniBand (NDR/XDR), Subnet Managers, and NVMe Software-Defined Storage.
- Certifications (Preferred): Red Hat OpenShift Specialist, CKA/CKS, or NVIDIA Certified Professional (InfiniBand / AI Infrastructure).
(Note: This is an intensive, hardware-facing infrastructure, networking, and systems architecture role)
Apply directly here on LinkedIn or send your CV to yashica@discovered.ae to start a confidential conversation with us.
Skills
Sign in to read and join the discussion.