AI Storage Engineer

Architecture & Design • Architect end-to-end storage solutions for large-scale AI training and inference clusters (NVIDIA DGX/HGX, SuperPOD, B300, GB300) • Design tiered data pipelines: ingest → NVMe hot tier → warm object → cold/archive • Integrate storage with GPUDirect Storage (GDS), NFS over RDMA, and InfiniBand/RoCE fabrics • Produce HLDs, LLDs, BoMs and technical documentation to enterprise standards Deployment & Operations • Deploy and manage parallel/distributed file systems — WEKA, VAST Data, DDN EXAScaler, Scale (GPFS), Lustre, BeeGFS • Configure multi-tenant storage for sovereign cloud and managed AI services • Implement checkpointing, snapshot, replication and DR strategies for AI workloads Performance & Benchmarking • Benchmark storage throughput, IOPS and latency using IO500, MLPerf Storage and fio • Tune systems for LLM training checkpoints, RAG pipelines and inference serving • Capacity planning, growth modelling and cost-per-TB optimisation Vendor & Stakeholder Management • Lead vendor evaluations, PoCs and commercial negotiations with storage OEMs • Partner with NVIDIA, VAST, WEKA, DDN, Pure Storage and Dell technical teams • Present designs and progress to internal leadership and enterprise customers REQUIRED QUALIFICATIONS • Experience: 7+ years in enterprise storage, with a minimum of 2 years in AI, HPC or GPU-cluster environments • File Systems: Deep expertise in one or more of WEKA, VAST Data, DDN, GPFS, Lustre, BeeGFS • Protocols: GPUDirect Storage, NFS over RDMA, SMB, S3, NVMe-oF • Networking: Working knowledge of InfiniBand (NDR/HDR) and RoCEv2 storage fabrics • Platforms: Hands-on with NVIDIA reference architectures (DGX, HGX, SuperPOD, BasePOD) • Automation: Linux administration, Python/Bash scripting, Ansible or Terraform • Soft Skills: Strong documentation, stakeholder communication and vendor management PREFERRED QUALIFICATIONS • Vendor certifications from VAST DATA, WEKA, DDN, or Pure Storage • NVIDIA-Certified Associate — AI Infrastructure & Operations • Experience with multi-tenant sovereign cloud or AI-as-a-Service platforms • Exposure to Kubernetes CSI drivers, Slurm, and Run:ai orchestration • Familiarity with data governance, encryption at rest, and cross-site replication • Prior experience deploying at Tier III+ data centres in the Middle East or APAC

Experience: 7+ years in enterprise storage, with a minimum of 2 years in AI, HPC or GPU-cluster environments. File Systems: Deep expertise in one or more of WEKA, VAST Data, DDN, GPFS, Lustre, BeeGFS. Protocols: GPUDirect Storage, NFS over RDMA, SMB, S3, NVMe-oF. Networking: Working knowledge of InfiniBand (NDR/HDR) and RoCEv2 storage fabrics. Platforms: Hands-on with NVIDIA reference architectures (DGX, HGX, SuperPOD, BasePOD).Automation: Linux administration, Python/Bash scripting, Ansible or Terraform. Soft Skills: Strong documentation, stakeholder communication and vendor management.

Architect end-to-end storage solutions for large-scale AI training and inference clusters (NVIDIA DGX/HGX, SuperPOD, B300, GB300). Design tiered data pipelines: ingest → NVMe hot tier → warm object → cold/archive. Integrate storage with GPUDirect Storage (GDS), NFS over RDMA, and InfiniBand/RoCE fabrics. Produce HLDs, LLDs, BoMs and technical documentation to enterprise standards. Deploy and manage parallel/distributed file systems — WEKA, VAST Data, DDN EXAScaler, Scale (GPFS), Lustre, BeeGFS. Configure multi-tenant storage for sovereign cloud and managed AI services. Implement checkpointing, snapshot, replication and DR strategies for AI workloads. Benchmark storage throughput, IOPS and latency using IO500, MLPerf Storage and fio. Tune systems for LLM training checkpoints, RAG pipelines and inference serving. Capacity planning, growth modelling and cost-per-TB optimisation. Lead vendor evaluations, PoCs and commercial negotiations with storage OEMs. Partner with NVIDIA, VAST, WEKA, DDN, Pure Storage and Dell technical teams. Present designs and progress to internal leadership and enterprise customers.