Principal Performance Architect

Leuven; Heverlee; Sur place; EUR 120 000 - 180 000; Plein temps

Résumé du poste

imec's Compute System Architecture Unit (CSA) in Leuven is seeking a Principal AI Performance Architect to drive software-hardware codesign for next-generation AI and HPC systems. You will lead system modeling, performance analysis, and projections across the stack, coordinating with experts to identify codesign solutions and build the infrastructure to evaluate them.

This role shapes future supercomputer architectures and collaborates across teams to meet product milestones and deliver

Responsabilités

• Research various solutions to enable the next large-scale HPC clusters.
• Develop software/hardware performance models to identify bottlenecks and propose improvements.
• Collaborate across teams to capture AI in HPC characteristics and explore design space for higher performance.
• Define and scope system and SoC requirements for future architectures.

Description du poste

What you will do

You will architect software-hardware codesign innovations for next-gen artificial intelligence high-performance computing systems.

The Compute System Architecture Unit (CSA) at imec is researching next-generation large-scale heterogeneous computer architectures. The team is responsible for workload characterization, runtime management innovations, performance modeling, architecture simulation and prototyping for future applications and the future systems to execute them, to reach multiple orders of magnitude better performance, energy-efficiency, and total-cost-of-ownership.

We are looking for a Principal AI Performance Architect to technically drive innovation across layers of the computing stack, from application to semiconductor technology innovation and everything in between. You work closely with experts from different fields to identify codesign solutions and to build the infrastructure required to evaluate them. As an architect, you will drive system modeling, performance analysis, projections, and strategic innovations against critical product life cycle milestones and program deliverables. This position plays a key role in our mission to shape the future of supercomputer architecture.

• Research various solutions to enable the next large-scale-high-performance computing clusters. The role spans over various layers from workload characterization, resource management, system architecture, and microarchitecture.
• Develop software/hardware performance models at various levels of detail to identify performance bottlenecks and propose new solution alleviating the bottlenecks.
• Work with multiple teams to capture the relevant characteristics of the use of AI in HPC applications, and to explore the design space for higher accuracy and higher performance of future HPC applications.
• Collaborate across the company to guide the direction of HPC HW/SW codesign based on the top HW issues and SW performance limiters generated from the models and characteristics.
• Define and scope system and SoC requirements for future architectures.
• Model features and configurations that improve performance, power, total cost of ownership and emissions.