Eugene de Beste

Multi-region GPU cloud (OpenStack-based). Title updated during internal restructuring; scope unchanged.

• Built the infrastructure operations function around a clear operating model, escalation paths, and permission/IAM scopes, separating L1/L2 support from infrastructure engineering and reducing repeat escalations into the engineering team.
• Owned observability platform strategy: designed a unified monitoring architecture feeding a new Network Operations Centre (NOC) and led build-vs-buy / total-cost-of-ownership selection (open-source Prometheus/VictoriaMetrics + DCGM vs commercial), scaling the approach toward a large-scale NVIDIA B200 SuperPOD region.
• Primary engineer for centralised bare-metal observability, building a NetBox-driven stack where in-region collectors feed a central VictoriaMetrics and Grafana deployment with an alert suite tuned for signal over noise. Currently centralising monitoring across the EU bare-metal region (two clusters).
• Built CX and L2 enablement across OpenStack, Linux, and networking: training tracks, runbooks, decision trees, and scoped self-service workflows.
• Coordinated data-centre, hardware, and partner engagement, and led the observability procurement process through vendor evaluation, scenario presentation, and partner justification.
• Established incident-response and root-cause-analysis (RCA) practice; led major-incident response and authored RCAs for customer-impacting outages.

• Led the OpenStack Research Cloud and ACE Lab at the CHPC, part of the CSIR, as R&D platforms for cloud/HPC experimentation, while also contributing to Sebowa production-cloud operations serving hundreds of researchers.
• Deployed the CHPC Pretoria region for Sebowa and trained the local team on OpenStack operations, support, and policy alignment.
• Architected and operated multi-petabyte Ceph storage for HPC workloads, building monitoring and inventory systems that improved operational visibility.
• Co-led the South African Student Cluster Competition programme, mentoring teams and developing training material for cohorts progressing to the ISC Student Cluster Challenge.

• Planned and delivered a private OpenStack cloud for astronomy and bioinformatics with multi-petabyte Ceph storage.
• Prototyped the platform on test hardware, then moved it into production with OpenStack, Ceph, and Manila evaluation for file services.
• Supported handover and ongoing technical guidance during deployment, including cloud and storage troubleshooting.

• Trained and supported student teams for the CHPC Student Cluster Competition, redesigning delivery for fully remote operation during the pandemic.
• Migrated in-house VM management onto OpenStack and supported HISP, NICD, and UWC HPC deployments across cloud and bare-metal environments.
• Automated bare-metal and cloud builds with MAAS, PXE, Ansible, and Terraform, introducing FreeIPA, monitoring, and change-management practices.
• Supported Ceph storage, networking, and helpdesk operations across research and IT teams.