Senior Infrastructure Engineer — OpenStack for AI & HPC Clusters
SENIOR INFRASTRUCTURE ENGINEER — OPENSTACK FOR AI & HPC CLUSTERS
We are looking for a Senior Infrastructure Engineer to lead our Infrastructure Team as its technical lead, owning the cloud platform that powers some of the largest GPU-accelerated AI and HPC clusters in Europe. The platform is OpenStack on bare metal (Kayobe / Kolla-Ansible), with some clusters delivered as Kubernetes-native bare metal; its tenants are Slurm partitions, distributed training jobs, and latency-sensitive inference services, consuming thousands of GPUs over RDMA fabrics and parallel storage.
We hire first for depth of systems engineering — the ability to understand why something does not work, not just how to restart it: reading an strace, reading the source of a service, isolating a kernel, driver, or network issue, and driving a fix upstream when needed. A strong engineer with this foundation who knows OpenStack — or has built comparable GPU / HPC infrastructure with other tools — is exactly who we are looking for; the specific stack can be learned, the way of reasoning cannot.
One thing we want to be explicit about: in this role you own the platform end to end, from the workload down to the physical server. If you have built and operated a platform yourself — provisioning the hardware, deploying the stack, and delivering it to its users — this is the role for you: broad competence across the whole deployment process matters more here than deep expertise in a single workload running on a cloud managed by another team.
KEY RESPONSIBILITIES:
- Own the architecture and day-2 operations of OpenStack control planes serving GPU clusters of hundreds of nodes today, with a growth path to 1000+ nodes at upcoming public and private AI Factories: uptime, capacity, performance, and security KPIs.
- Operate the platform lifecycle on bare metal — from node provisioning (Ironic, out-of-band management, firmware baselines) to GPU delivery to tenants (PCI passthrough, MIG / vGPU partitioning on multi-GPU nodes).
- Build and operate Kubernetes for AI workloads — on OpenStack (Magnum / Cluster API) and Kubernetes-native on bare metal — including GPU operator and the platform side of inference serving stacks (e.g., NVIDIA Triton / NIM, vLLM).
- Engineer the tenant networking that lets AI workloads reach the high-performance fabric at line rate (Neutron OVN / OVS, SR-IOV with RDMA towards InfiniBand /RoCE, BGP-EVPN), in partnership with the Network & Interconnect role, which owns the fabric design.
- Diagnose and resolve complex, cross-layer production issues down to the root cause — kernel, GPU driver, storage, and network stack — reading service source code where needed and contributing fixes upstream; manage escalations with the vendors behind the platform (NVIDIA, Dell, storage and fabric vendors).
- Develop and maintain Infrastructure-as-Code pipelines (Ansible, OpenTofu / Terraform, Helm)
with gitops-style review for production change management; maintain operational runbooks — with explicit go / no-go gates and rollback plans — and RCA reports.
- Lead the Infrastructure Team day to day: mentor mid-level and junior engineers and run their performance reviews, own the internal lab and the team’s quality standards, review technical documents for internal and customer use, and align with the other team leaders.
- Collaborate with the Presales team in designing systems end-to-end from hardware configuration to software stack: translate tender requirements into buildable configurations and validate new solutions with PoCs and benchmarks.
EDUCATION:
Master’s degree or Ph.D. in Computer Science, Computational Science, Physics, Network Engineering, or a related STEM field — or equivalent practical experience, including:
- 5+ years of hands-on engineering experience in Linux systems, infrastructure automation, and complex production systems.
- 3+ years building and operating an OpenStack platform (or comparable bare-metal / GPU cloud infrastructure) as the platform team — designing, deploying, upgrading, and troubleshooting the platform itself, not consuming it as a tenant.
- 2+ years leading a technical team, as team lead or technical lead: mentoring, reviews, planning.
- Direct exposure to GPU, HPC, or AI infrastructure (GPU servers, RDMA fabrics, batch schedulers, parallel storage) is a strong differentiator.
SKILLS AND COMPETENCES - CORE TECHNICAL SKILLS
Linux & systems engineering (deep):
- Expert Linux sysadmin (Rocky / RHEL, Ubuntu): kernel, systemd, namespaces / cgroups, storage and network stack, GPU driver lifecycle; container runtimes and their internals.
- Advanced troubleshooting strace, perf, ftrace / eBPF, gdb; comfortable reading the source of the services operated and contributing fixes upstream;
OpenStack on bare metal
- Strong, hands-on proficiency on Nova, Neutron, Ironic, Cinder, Keystone, Placement; Ceph as the storage backbone.
- Familiarity with PXE boot and out-of-band management (Redfish / IPMI); direct experience with Kayobe + Kolla-Ansible is a strong plus.
Kubernetes for GPU workloads
- Production-grade Kubernetes on bare metal and on OpenStack; GPU operator and device plugins; familiarity with inference serving stacks (NVIDIA Triton / NIM, vLLM) is a plus.
AI / HPC workload literacy
- Working understanding of what the platform serves — batch training as well as online inference: Slurm,
NCCL and distributed training patterns, GPU partitioning (MIG / vGPU) — enough to reason about platform decisions from the workload’s point of view.
Networking
- OVN / OVS, SR-IOV, RDMA over InfiniBand / RoCE; working knowledge of EVPN / VXLAN and spine-leaf datacenter fabrics (fabric design is owned by the Network & Interconnect role).
Automation & GitOps
- Production-quality automation in Bash, Python, or Go; hands-on Ansible, Terraform / OpenTofu, Helm.
- Solid Git operations and CI pipelines (GitLab CI or equivalent) for production change management.
Communication
- Fluent English (working language with NVIDIA / Mellanox / Dell engineering); clear written and verbal communication for vendor, customer, and internal audiences.
SKILLS AND COMPETENCES - SOFT SKILLS
- Technical Leadership — Ability to guide technical decisions and mentor engineers.
- Problem Solving — Strong analytical skills to resolve complex, cross-layer infrastructure issues.
- Collaboration — Ability to work effectively with internal teams and external partners.
- Ownership — Strong sense of responsibility and accountability for systems and outcomes.
- Decision Making — Ability to act decisively in high-pressure, production-critical environments.
- Adaptability — Comfortable working in fast-evolving, complex infrastructure environments.
- Knowledge Sharing — Commitment to mentoring and promoting best practices within the team.
WORKING ENVIRONMENT:
- Remote work with travel — Work from anywhere in Italy, with no hybrid mandate. Expect to travel a few days a month, as the role requires: periodic group meetings and company events, on-site engagements at customer and partner facilities in Italy and across the EU (cluster bring-up, vendor PoC, co-deployment), and national and international conferences. Travel is covered by the company.
- Objective-based work — Clear, measurable objectives reviewed and updated throughout the year, with a structured growth path and access to leading-edge HPC and AI infrastructure in production.
- RAL: 45.000,00 € - 65.000,00 €
GROWTH & DEVELOPMENT
While this position is targeted at senior engineers, we also welcome applications from mid-level engineers with a solid systems-engineering foundation — strong Linux fundamentals and hands-on experience building (not just using) OpenStack, Kubernetes, or large-scale GPU / HPC infrastructure — looking to grow into full ownership of a production AI cloud platform. What matters most is depth of reasoning and a genuine curiosity for how systems work under the hood; the specific stack is something we help you master.
Depending on experience level, successful candidates will grow into increased responsibility through hands-on work on large-scale AI and HPC infrastructures, supported by structured mentorship and knowledge sharing within the team.
We are looking for a Senior Infrastructure Engineer to lead our Infrastructure Team as its technical lead, owning the cloud platform that powers some of the largest GPU-accelerated AI and HPC clusters in Europe. The platform is OpenStack on bare metal (Kayobe / Kolla-Ansible), with some clusters delivered as Kubernetes-native bare metal; its tenants are Slurm partitions, distributed training jobs, and latency-sensitive inference services, consuming thousands of GPUs over RDMA fabrics and parallel storage.
We hire first for depth of systems engineering — the ability to understand why something does not work, not just how to restart it: reading an strace, reading the source of a service, isolating a kernel, driver, or network issue, and driving a fix upstream when needed. A strong engineer with this foundation who knows OpenStack — or has built comparable GPU / HPC infrastructure with other tools — is exactly who we are looking for; the specific stack can be learned, the way of reasoning cannot.
One thing we want to be explicit about: in this role you own the platform end to end, from the workload down to the physical server. If you have built and operated a platform yourself — provisioning the hardware, deploying the stack, and delivering it to its users — this is the role for you: broad competence across the whole deployment process matters more here than deep expertise in a single workload running on a cloud managed by another team.
KEY RESPONSIBILITIES:
- Own the architecture and day-2 operations of OpenStack control planes serving GPU clusters of hundreds of nodes today, with a growth path to 1000+ nodes at upcoming public and private AI Factories: uptime, capacity, performance, and security KPIs.
- Operate the platform lifecycle on bare metal — from node provisioning (Ironic, out-of-band management, firmware baselines) to GPU delivery to tenants (PCI passthrough, MIG / vGPU partitioning on multi-GPU nodes).
- Build and operate Kubernetes for AI workloads — on OpenStack (Magnum / Cluster API) and Kubernetes-native on bare metal — including GPU operator and the platform side of inference serving stacks (e.g., NVIDIA Triton / NIM, vLLM).
- Engineer the tenant networking that lets AI workloads reach the high-performance fabric at line rate (Neutron OVN / OVS, SR-IOV with RDMA towards InfiniBand /RoCE, BGP-EVPN), in partnership with the Network & Interconnect role, which owns the fabric design.
- Diagnose and resolve complex, cross-layer production issues down to the root cause — kernel, GPU driver, storage, and network stack — reading service source code where needed and contributing fixes upstream; manage escalations with the vendors behind the platform (NVIDIA, Dell, storage and fabric vendors).
- Develop and maintain Infrastructure-as-Code pipelines (Ansible, OpenTofu / Terraform, Helm)
with gitops-style review for production change management; maintain operational runbooks — with explicit go / no-go gates and rollback plans — and RCA reports.
- Lead the Infrastructure Team day to day: mentor mid-level and junior engineers and run their performance reviews, own the internal lab and the team’s quality standards, review technical documents for internal and customer use, and align with the other team leaders.
- Collaborate with the Presales team in designing systems end-to-end from hardware configuration to software stack: translate tender requirements into buildable configurations and validate new solutions with PoCs and benchmarks.
EDUCATION:
Master’s degree or Ph.D. in Computer Science, Computational Science, Physics, Network Engineering, or a related STEM field — or equivalent practical experience, including:
- 5+ years of hands-on engineering experience in Linux systems, infrastructure automation, and complex production systems.
- 3+ years building and operating an OpenStack platform (or comparable bare-metal / GPU cloud infrastructure) as the platform team — designing, deploying, upgrading, and troubleshooting the platform itself, not consuming it as a tenant.
- 2+ years leading a technical team, as team lead or technical lead: mentoring, reviews, planning.
- Direct exposure to GPU, HPC, or AI infrastructure (GPU servers, RDMA fabrics, batch schedulers, parallel storage) is a strong differentiator.
SKILLS AND COMPETENCES - CORE TECHNICAL SKILLS
Linux & systems engineering (deep):
- Expert Linux sysadmin (Rocky / RHEL, Ubuntu): kernel, systemd, namespaces / cgroups, storage and network stack, GPU driver lifecycle; container runtimes and their internals.
- Advanced troubleshooting strace, perf, ftrace / eBPF, gdb; comfortable reading the source of the services operated and contributing fixes upstream;
OpenStack on bare metal
- Strong, hands-on proficiency on Nova, Neutron, Ironic, Cinder, Keystone, Placement; Ceph as the storage backbone.
- Familiarity with PXE boot and out-of-band management (Redfish / IPMI); direct experience with Kayobe + Kolla-Ansible is a strong plus.
Kubernetes for GPU workloads
- Production-grade Kubernetes on bare metal and on OpenStack; GPU operator and device plugins; familiarity with inference serving stacks (NVIDIA Triton / NIM, vLLM) is a plus.
AI / HPC workload literacy
- Working understanding of what the platform serves — batch training as well as online inference: Slurm,
NCCL and distributed training patterns, GPU partitioning (MIG / vGPU) — enough to reason about platform decisions from the workload’s point of view.
Networking
- OVN / OVS, SR-IOV, RDMA over InfiniBand / RoCE; working knowledge of EVPN / VXLAN and spine-leaf datacenter fabrics (fabric design is owned by the Network & Interconnect role).
Automation & GitOps
- Production-quality automation in Bash, Python, or Go; hands-on Ansible, Terraform / OpenTofu, Helm.
- Solid Git operations and CI pipelines (GitLab CI or equivalent) for production change management.
Communication
- Fluent English (working language with NVIDIA / Mellanox / Dell engineering); clear written and verbal communication for vendor, customer, and internal audiences.
SKILLS AND COMPETENCES - SOFT SKILLS
- Technical Leadership — Ability to guide technical decisions and mentor engineers.
- Problem Solving — Strong analytical skills to resolve complex, cross-layer infrastructure issues.
- Collaboration — Ability to work effectively with internal teams and external partners.
- Ownership — Strong sense of responsibility and accountability for systems and outcomes.
- Decision Making — Ability to act decisively in high-pressure, production-critical environments.
- Adaptability — Comfortable working in fast-evolving, complex infrastructure environments.
- Knowledge Sharing — Commitment to mentoring and promoting best practices within the team.
WORKING ENVIRONMENT:
- Remote work with travel — Work from anywhere in Italy, with no hybrid mandate. Expect to travel a few days a month, as the role requires: periodic group meetings and company events, on-site engagements at customer and partner facilities in Italy and across the EU (cluster bring-up, vendor PoC, co-deployment), and national and international conferences. Travel is covered by the company.
- Objective-based work — Clear, measurable objectives reviewed and updated throughout the year, with a structured growth path and access to leading-edge HPC and AI infrastructure in production.
- RAL: 45.000,00 € - 65.000,00 €
GROWTH & DEVELOPMENT
While this position is targeted at senior engineers, we also welcome applications from mid-level engineers with a solid systems-engineering foundation — strong Linux fundamentals and hands-on experience building (not just using) OpenStack, Kubernetes, or large-scale GPU / HPC infrastructure — looking to grow into full ownership of a production AI cloud platform. What matters most is depth of reasoning and a genuine curiosity for how systems work under the hood; the specific stack is something we help you master.
Depending on experience level, successful candidates will grow into increased responsibility through hands-on work on large-scale AI and HPC infrastructures, supported by structured mentorship and knowledge sharing within the team.