Become part of our unique Performance Engineering Team and push run-time and memory efficiency of embedded GPU software to its limit. Join us in our task to reduce the run-time and memory footprint of GPU based software, while keeping it readable, maintainable, numerically stable and safe.
- You optimize GPU applications - in terms of throughput, latency or memory bandwidth.
- With a high degree of autonomy, you get the chance to shape large parts of your work, from the identification and analysis of hotspots to the delivery of the optimized result to the (internal) customer.
- In select cases, you will transform existing C++ software from the CPU to highly efficient GPU implementations.
- A vital part of the job are performance improvements for AI / inference.
- A high variety in functional domains and embedded targets allow insights into an exceptionally broad array of interesting topics to discover, comprehend and master.
To optimize GPU code, you rely on your knowledge and experience regarding:
- Implementation of high-quality, robust & maintainable massively parallel algorithms for GPUs.
- GPU development frameworks (vendor specific, e.g. nSight) for the creation & analysis of programs down to the assembly level.
- Modern GPU architectures and their properties including memory layouts, interconnections etc.
- Transformation from source code into machine code by the compiler.
- Mathematical reasoning.
- Education: M.Sc. or PhD in Computer Science or similar (university degree)
- Personality and Working Practice: easy adjustment to new environments, at times in a high frequency, "try things out", but in a systematic way with a clear prior goal and an explicit hypothesis what the outcome might be, patiently understand our customers goals and restrictions, and see them as an intriguing challenge
- Experience and Knowledge: several years of experience with GPU programming, very good knowledge in modern GPU architectures, strong mathematical background, ability to analyze and debug GPU related performance issues, CPU performance engineering knowledge is a plus
- Languages: fluent in verbal and written English, German is a plus
Apply now in just 3 minutes!
You want to work remotely or in part-time - we offer great possibilities for mobile working and different part-time models.
Need support during your application?
Miriam Steimmig (Human Resources)
+49 711 811 27525
Need further information about the job?
Christian Hütter (Functional Department)
+49 711 811 32653