Jose E Torres

Senior HPC Architect and DevOps Engineer with a diverse skill set developed through projects in both HPC and emerging cloud technologies.

• Led the deployment of a Canonical Charmed OpenStack private cloud (MAAS + Juju + OpenStack), enabling scalable internal cloud services (HPC + Kubernetes cluster) for research departments.

• ARGOCD and KARGO evaluation: Evaluation of functionalities on the ARGOCD CI/CD tool and provided a POC for KARGO.

• Evaluated and tested Kubernetes Kueue scheduler for AWS EKS, testing the resource allocation for HPC/ML/AI workloads.

• Integrated CERN's CernVM-FS into AWS EKS using a custom CSI driver, giving researchers inmediate access to scientific software stacks from Kubernetes pods.

• Designed and delivered an HPC cloud-bursting solution for a Synchrotron Light facility on GCP, enabling dynamic scale-out of compute resources for the HPC on-premise solution.

• Participated in the development of a hybrid workflow management PAAS, enabling researchers to submit hybrid workflows (HPC/ML/AI) in an energy-efficient manner.

• Assisted in the deployment of a Spanish national supercomputer, automating provisioning and configuration with Cobbler and Saltstack.

• Supported lift-and-shift migration of HPC workloads to AWS (Terraform, Ansible, CI/CD, Lustre), maintaining SLAs for critical research workflows.

Linux System administrator at the Operations team, responsible for maintaining the IT infrastructure of the center, including general services and the supercomputing environment.

The HPC environment was available for internal research and external institutions like RES (Supercomputing Spanish Network) and PRACE.

Participated in projects like:

• Responsible for the installation and configuration of a new highly available DNS system based on BIND, managing the main domain of the center and configuring some security measures like chroot, TSIG keys, and zone visibility restrictions.

• Responsible for the analysis, design and implementation of a new communication service based on RocketChat + LDAP. Improving the communication mechanism for the entire center and providing newer media functionalities for the users and departments.

• Responsible for the design and implementation of an HA virtualization environment based on KVM (oVirt), providing a more resilient and simple way to manage some of the main IT services of the center, and allowing the users and the research departments to request and deploy custom VMs for their uses.

• Responsible for the migration and integration of the team in charge of the PhylomeDB (http://

service, migrating their VMs and huge SQL DBs from another center.

The work included some tunings on the DBs based on MariaDB to improve their performance.

• Promoter for a DevOps approach on the team, integrating some methodologies and tools like Ansible, Gitlab usage, Kanban and CI/CD, simplifying and reducing the overhead for some recurrent daily procedures.

• Responsible for the analysis and performance improvement of the main backup system of the center based on TSM, reducing overlappings between tasks and restructuring the backup planning.

• Maintained VMware-based virtualization infrastructure, supporting multiple departments.

• Executed disaster recovery procedures for VM backup/restore.

• Automated infrastructure checks and updates with PowerShell scripts.

• Tested VMware vMotion performance during company-wide infrastructure integration.