Principal Network Architect

Job Description: Principal Network Architect (AI Fabric & Multi-Vendor SDN)
*Location:* UAE (Dubai / Abu Dhabi)
*Position Type:* Full-Time, Permanent
*Domain:* Enterprise AI Infrastructure, Sovereign AI Cloud & Hyperscale Networking
---
Role Overview
As *Principal Network Architect*, you will serve as the technical authority spearheading the end-to-end strategy, architecture, design, and deployment of mission-critical networking infrastructure across the UAE. This role sits at the intersection of ultra-high-performance AI compute fabrics and enterprise multi-domain networks.
You will lead the engineering of low-latency, lossless back-end fabrics powering large-scale GPU clusters (LLM training and inference), federated multi-domain enterprise data centers, and compliant *Sovereign AI network ecosystems* adhering strictly to UAE national cybersecurity and data residency mandates.
---
Core Responsibilities
AI Fabric & Ultra-Low-Latency Architecture
* Architect, benchmark, and deploy non-blocking, line-rate back-end networks for distributed AI training/inference clusters using both *InfiniBand* (NDR/XDR, Quantum switches, Subnet Management, adaptive routing) and *Lossless RoCEv2* (PFC, ECN, DCQCN, ETS, buffer tuning, PFC deadlock prevention).
* Engineer front-end and out-of-band management fabrics connecting heterogeneous compute, high-throughput parallel file storage (GPFS, Lustre, NVMe-oF), and client-facing endpoints.
* Design telemetry and observability pipelines tailored for AI fabrics using streaming telemetry, gNMI, and in-band network telemetry (INT) to eliminate GPU tail-latency, silent packet drops, and fabric bottlenecks.
Multi-OEM & Multi-Domain SDN Design
* Direct architectural strategy across multi-vendor ecosystems (NVIDIA/Mellanox, Cisco, Arista, Juniper/Mist, Huawei CloudEngine, Extreme).
* Architect automated multi-tenant overlay/underlay networks utilizing *EVPN-VXLAN*, Segment Routing (SR-MPLS / SRv6), and BGP across Data Center Interconnect (DCI), campus, edge, and sovereign private clouds.
* Drive SDN programmability and infrastructure-as-code (IaC) workflows using Terraform, Ansible, Python, and CI/CD pipelines to achieve automated zero-touch provisioning (ZTP).
Sovereign AI & Regulatory Compliance
* Architect secure, air-gapped, and isolated sovereign cloud network zones meeting UAE-specific data sovereignty, security, and governance standards (e.g., DESC, TDRA, UAE Cyber Security Council, national AI governance frameworks).
* Design robust zero-trust network architectures (ZTNA), post-quantum cryptographically resilient interconnects, and strict perimeter controls protecting sovereign data assets and national model training pipelines.
Cross-Domain Leadership & Advisory
* Bridge domain silos by harmonizing network architecture with enterprise systems: High-Performance Computing (HPC), GPU infrastructure, storage area networks (SAN/NAS), virtualization, container orchestration (Kubernetes, CNI plugins like Multus/SR-IOV), and multi-tenant cloud operations.
* Serve as the principal technical advisor for C-level stakeholders, national entities, and strategic enterprise partners across the UAE and broader GCC region.
---
Qualifications & Experience
* *Experience:* 12+ years in high-tier network architecture, engineering, and deployment, with at least 4+ years dedicated to AI fabrics, HPC interconnects, or hyperscale cloud environments.
* *Fabric Expertise:* Mastery of InfiniBand architectures (NVIDIA Quantum platform, OpenSM tuning) alongside RoCEv2 deployments across leaf-spine / rail-optimized topologies.
* *Multi-OEM Mastery:* Hands-on architectural and validation experience across at least two major networking OEMs (e.g., NVIDIA Mellanox + Arista EOS, Cisco Nexus/ACI, or Juniper Junos).
* *Protocols & Standards:* Deep mastery of BGP, EVPN-VXLAN, SRv6, RoCEv2 congestion management mechanisms, QoS/traffic class shaping, and multi-tier spine-and-leaf fabrics.
* *Automation:* Proficiency in network automation frameworks (Ansible, Terraform, Netmiko/NAPALM, GitOps) and scripting in Python or Go.
* *Regional Regulatory Awareness:* Demonstrable understanding of UAE regulatory environments, data residency standards, and sovereign infrastructure deployment models.
Preferred Certifications & Credentials
* CCIE / CCDE (Data Center or Enterprise Infrastructure)
* Arista Cloud Engineer (ACE 5/6)
* NVIDIA Certified Associate or Professional (Cumulus / InfiniBand / AI Networking)
* JNCIE-DC or equivalent vendor-neutral hyperscale credentials

Work arrangement
No