Talk to me about Sovereign Cloud, GPUaaS, LLM inference infrastructure, OpenStack, Ceph, Kubernetes, InfiniBand and GPU storage. I design and build sovereign cloud and GPU-as-a-Service platforms, from bare metal to tenant-ready services. For 12+ years I’ve delivered solutions and support for highly available, secure infrastructure that runs mission-critical systems. My work spans the full stack: MAAS-based bare-metal provisioning, OpenStack private cloud deployments, Ceph and enterprise storage, Kubernetes platforms, network fabrics, and InfiniBand fabrics for GPU clusters, including high-throughput storage such as pNFS for AI/ML training and LLM inference workloads. I’ve led many migrations, cloud deployments, infrastructure provisioning and toolset rollouts, and run storage POCs across vendors to find the right fit for each requirement. I’ve managed both private (VMware vCloud Director, OpenStack) and public (AWS, Azure) cloud environments, and supported leading organisations including Alstom (Transport, UK), Circle K (Fuel, Europe), CPA Global (Data Analytics, USA), Royal Mail Group (Courier, UK) and Snapdeal. Skills • Cloud & GPU: Sovereign cloud, GPUaaS, LLM inference infrastructure, OpenStack, MAAS, Kubernetes • Storage: Ceph, pNFS / GPU storage, NetApp, IBM Storwize, Nexenta • Networking: Data centre network fabrics, InfiniBand • Virtualization: VMware, RHEV, KVM • Public cloud: AWS, Azure • Automation: Ansible, Ansible Tower, Jenkins, Bash scripting • High availability: Pacemaker (PCS), HAProxy, Keepalived, Luci/Ricci • Hardware: Dell, HPE, Cisco UCS, Ericsson HDS 8000 • Operating systems: RHEL, CentOS, Ubuntu, SUSE, Windows, Kali Linux • Monitoring: Prometheus, Grafana, Telegraf, InfluxDB, Kibana • Team management, ITIL, solution workshops I can help you with • Designing a GPUaaS or AI cloud: compute, fabric, storage and the multi-tenancy layer • OpenStack solutioning and design: architecture, service selection, control plane HA, networking (Neutron/OVN), sizing, and upgrade or migration planning • Ceph solutioning and design: cluster sizing, hardware selection, replication vs. erasure coding, performance tuning, and integration with OpenStack and Kubernetes • Infrastructure for LLM inference and training at scale • Sovereign / private cloud architecture and build-vs-buy decisions • Storage choices for GPU workloads and vendor POC evaluation • InfiniBand vs. Ethernet (RoCE) fabric decisions • Reviewing your solution design, BoM or RFP response before it goes out • Career guidance for engineers moving into cloud or solution architecture roles