We invite applications for a PhD position in the Lift team. This position is fully funded (for EU nationals) by a scholarship from the School of Computing Science at the University of Glasgow. The starting date is flexible but should be before the 1st January 2019. Deadline for the application is the 31st of July 2018.
Applications based on artificial intelligence continue to profoundly change the ways computer systems and humans interact with each other. To further improve computers' understanding and enhance the intelligence of machines, we build increasingly complex algorithms and models which require more and more powerful computer hardware. To accommodate the increasing demand for higher performance and efficiency of new applications, the hardware landscape is specializing and diversifying faster than ever. This offers exciting opportunities as novel hardware architectures offer greater efficiency enabling applications which were unthinkable just a few years ago. Unfortunately, with current programming languages and tools each new hardware target requires costly re-writing and re-optimization of the software. Instead, we would like to write portable software which achieves high performance on each target hardware.
Lift (www.lift-project.org) is a promising novel approach aiming to achieve this performance portability on modern parallel architectures via a sophisticated optimizing compiler which explores optimization opportunities expressed as formal rewrite rules. This PhD project will give you the opportunity to contribute to the Lift project and help shaping the future of parallel programming and the optimization of machine learning applications.
We are interested in build upon and strengthen the core technologies developed in the Lift project. Depending on your strength, interests, and motivation there are different possible directions for ranging from 1) foundational work in programming language theory which underpins the Lift language and compiler to 2) applied work optimizing real-world computer vision and machine learning applications and 3) enabling Lift to target specific hardware architectures such as FPGAs.
You will join the vibrant Lift research team which by September 2018 will consists of 8 PhD students, one post doctoral researcher, and two leading academics - all born in 8 different countries. PhD students from the Lift team have been engaging with our industrial partners via internships at Oracle Labs, Codeplay, Microsoft Research, Nvidia, and Lawrence Livermore National Laboratory.
Candidates interested by this opportunity should first contact Michel as soon as possible by email (firstname.lastname@example.org) attaching: a CV; transcripts (courses and grades achieved during BSc / MSc); and a short statement explaining how your background is relevant for Lift.
By the 31th of July you will have to completed the University of Glasgow PhD application process: https://www.gla.ac.uk/research/opportunities/howtoapplyforaresearchdegree/
More information can also be found here.
The ideal starting date would be in the autumn of 2018, but this is flexible. You should be able to start at latest by 1st Janurary 2019.
Ideal candidates will have an excellent degree (BSc or MSc) in computer science or related discipline. You should have a strong interest in programming languages and compilers for modern parallel hardware architectures.
The ideal candidate would have the following skills:
- Good skills in functional programming or good skills in parallel programming using OpenCL or CUDA
- Knowledge and strong interest in compiler design or programming language design
- Basic understanding of machine learning algorithms
- Excellent collaboration and communication skills
- Strong motivation to succeed within a competitive research field
- Flexibility and ability to work alone and in a team environment
- Fluent in spoken and written English
These are full UK/EU scholarships, covering the UK/EU fee and a standard stipend for the duration of a 3 year PhD. However, we will also consider international applicants from outside the EU, if they demonstrate that they can cover the remaining fees from other sources.