Research Scientist (m/f/x) Next-Generation Computing Architectures Fixed Term (2 Years)

Description

The company is ZEISS Corporate Research & Technology, which looks at scientific and technological frontiers to spot emerging computing architectures and turn them into future technology opportunities for the company.

The team is looking for a research scientist who knows accelerator architectures and new computing concepts shaped by recent AI progress. That covers CPUs, GPUs and TPUs, custom accelerators, optical and analog computing, new memory designs and novel forms of heterogeneous computing.

The role mixes strategic technology exploration with practical technical work. You will build focused proof-of-concepts to test promising architecture ideas, and you will talk directly with startups, research groups and technology partners to see their solutions up close, judge how mature they are and find the best options for the company.

Responsibilities

• Analyze AI computing systems from end to end, looking at how training and inference workloads run on modern data-center hardware and pinpointing core bottlenecks in compute, memory, interconnects and energy use.
• Evaluate established, emerging and radically new architectures, such as custom accelerators, chiplets, in-memory and near-memory computing, optical and analog computing, and heterogeneous systems.
• Turn architecture hypotheses into hands-on demonstrators and measure their potential against the current state of the art.
• Work with startups, research groups and technology providers to judge maturity and differentiation, and separate real technology shifts from incremental gains.

Requirements

• A university degree in computer science, computer engineering, electrical engineering (with a focus on digital systems or computer architecture) or a closely related field. A PhD or equivalent research experience is welcome.
• Expertise in at least two of: computer architecture and systems, machine-learning systems and AI infrastructure, heterogeneous accelerators and emerging computing architectures.
• A systems-level grasp of modern AI stacks, including training and inference workflows, numerical kernels, memory hierarchies, data movement, interconnects, accelerator hardware and system software. You must be able to explain where bottlenecks arise, why they arise, and which architectural approaches address them.
• Programming skills in Python and/or C/C++. Experience with GPU computing, parallel programming, distributed systems, compilers, runtimes or HPC environments is an advantage.
• Curiosity about external technology developments and the ability to talk with startups, research groups and technology providers, understand their approaches in technical depth and assess how relevant they are.
• An analytical view of technology trends, willingness to change your view when the evidence changes, and the ability to present complex topics clearly. Business-fluent English is required.

Conditions

This is a fixed-term position of 2 years.

Equal opportunity applies to all applicants regardless of background.