Associate Director - HPC Infrastructure Architect

About the role

Novartis' Cloud & Hosting Services has set up AI & HPC Engineering as the single channel for high-performance computing, and Architecture & Integration owns its target architecture. This Associate Director role is accountable for the compute, storage, interconnect, scheduling and data-locality design that scientific workloads use, on premises and in the cloud. It is an individual contributor architecture role in which credibility with scientists is as important as credibility with infrastructure engineers.

Key responsibilities

• Own the HPC reference architecture: cluster topologies, CPU and GPU node designs, storage, interconnect fabric, schedulers and the supporting software stack.
• Define workload placement: which workloads run on dedicated clusters, which burst to the cloud and which belong on AI platforms, based on performance, data gravity, cost and compliance.
• Produce and maintain capacity plans for compute, accelerators, storage and network, including power, rack and cooling requirements.
• Establish data-locality and data-movement principles, including replication and recovery for irreplaceable scientific data.
• Translate scientific requirements into sizing, benchmarking and procurement input, and validate feasibility before engineering build.
• Maintain architecture decision records linking each choice to scientific outcome, cost and lifecycle consequence.
• Define the interface with AI infrastructure so shared accelerators, storage and data pipelines are used coherently.
• Ensure designs meet security, quality and regulatory expectations and are operable by Infrastructure Operations after handover.
• Act as the trusted architecture contact for scientific communities and research IT partners, coach engineers and analysts, and embody the company's values.

Required profile

• 8+ years of experience designing, building or operating large-scale HPC environments.
• Proven expertise in HPC platform architecture, capacity planning and multi-cluster infrastructure design.
• Strong experience supporting Linux-based compute clusters in enterprise or scientific computing environments.
• Experience with high-performance storage, parallel file systems and high-speed networking, including InfiniBand and Ethernet fabrics.
• Familiarity with containerized, reproducible scientific workflows and hybrid or cloud-based HPC environments.
• Fluent English, written and spoken; additional languages are an asset.

Required skills

• Workload schedulers such as Slurm, LSF or comparable technologies.
• CPU and GPU infrastructure design, performance optimization and benchmarking.
• Data replication, archiving and infrastructure planning (power, cooling, rack design).

What we offer

• Expected annual base salary of 1,120,210 to 2,080,390 CZK, with a possible performance-based bonus.
• Benefits such as insurance plans, retirement plans, wellbeing resources and global recognition programs.
• Flexible and hybrid working where possible, and at least 14 weeks of paid parental leave.