— Applied Polyhedral Compilation —
Polly Labs is looking for a post-doctoral researcher (or senior engineers with equivalent qualifications) to work with us on applied polyhedral loop optimization. Our goal is to advance the practical impact of polyhedral loop modeling techniques. We invite you to join our team to push the boundaries of today’s loop optimization and accelerator mapping techniques. Using and extending the LLVM Polly loop optimization infrastructure, we jointly develop state-of-the-art loop optimization techniques. Increased scalability, better and faster mathematical models, jit-compilation, and new accelerators such as FPGAs are topics we are actively working on. As a postdoctoral researcher we expect you to both closely interact with us on these topics, but also to introduce and lead new research that expands the current state-of-the-art in the context of loop optimization and accelerator compilation.
We are looking for exceptional candidates and not perfect job-profile matches! If you are strong in math or programming and have an exciting academic or industrial background, don’t hesitate to apply!
- Expert in compilation and code generation (e.g., LLVM, gcc, …)
- Strong mathematical background (e.g., linear programming, (non)-linear optimization, …)
- First high-quality publications
- Experience with:
- Polyhedral loop modeling (e.g., Polly, isl, Pluto)
- Accelerators (e.g., CUDA, OpenCL, Vulkan)
- Machine Learning Frameworks and the underlying computations (Tensorflow, Torch, Caffee, …) (plus)
- Performance modeling and generation of performance models (plus)
- The use of machine learning and autotuning to derive good compiler heuristics (plus)
- Develop high-quality software and have experience with open source communities
- Have a strong personal research agenda which you develop and advance independently
- Enjoy to collaborate and interact with other researchers
- Shape and design your own research agenda independently (within the overall project goals)
- Work on practical large scale problems: we optimize the weather model that delivers daily forecasts in Germany, Switzerland, Japan, and many other countries; we also look into large-scale data-query optimization and machine learning.
- Get access to the worlds fastest supercomputer outside China
- Close collaboration with the open source community (e.g., LLVM, Polly, …)
- Close interaction with industry (e.g., with Xilinx, ARM, Qualcomm)
- Work in the center of Paris at Ecole Normale Superieure, France’s leading research university. Hosted by the PARKAS group, you are in daily interaction with world-leading compiler experts in the group of Albert Cohen.
- You regularly work, interact, and visit Polly Labs members at ETH Zurich and other Polly Labs partners.
This senior position requires a doctoral degree (or relevant experience) in computer science, computer engineering, or equivalent experience. A successful candidate has significant experience in compilation, parallel programming, and analytical modeling. A proven record of working and collaborating on larger research and development projects, and a record of developing high-quality (open source) software is important to us. We do not expect expertise in all areas, but believe that an excellent candidate adds expert knowledge in one of our core areas of interest.
Excellent skills in spoken and written English are required.
Living and working in Paris (France) at Ecole Normale Superieure
France, and especially Paris as its capital, is an exceptional place to live and work. The French ``Joie de vivre’’ – the joy of living – not only refers to exceptional food and wine, but really a unique lifestyle. The international touch of a global city, several world-known universities and research institutes, and a large English speaking expat community provide plenty of connection points even for non-french speakers.
Ecole Normale Superieure is France’s leading research university. Even though comparatively small, the university has educated ten Fields Medal and eight Nobel Prize recipients and is by far leading the Nobel Prize per alumni ranking worldwide. The ENS computer science department, hosts leading researchers in machine learning, synchronous systems, cryptography, abstract interpretation, and compilation and provides plenty of collaboration opportunities. ENS also works closely with IRILL, the french center for research and innovation on free software, and ENS researchers often take part in the Paris LLVM social events.
Your application should include a letter of motivation (at most 2 pages), covering your research interests and how they will support this project, a description of your PhD thesis (or relevant work experience), and evidence of independent self-motivated research and software development. Please highlight experience in collaborative projects and especially open source software development. Please include your CV, degrees and grades, a copy of your thesis, publications (including descriptions of the candidates contributions), other relevant documents, and the earliest possible starting date. Please also include at least one letter of recommendation and contact information for references.
Starting date: Anytime from now
Working Location: Paris, France – Ecole Normale Superieure
Please email your application to tobias dot grosser at inf dot ethz dot ch and include the tag [JOB] in the subject line.