About Navjot Kukreja

Hi. I am a Lecturer (Assistant Professor) in the Department of Computer Science at the University of Liverpool. Previously, I worked at Imperial College London.

Research Interests

My research is focused around High-performance computing, Numerical methods (PDE Solvers), machine learning and inverse problems. We are especially interested in Scientific Machine Learning and Physics-Informed Machine Learning. If you are interested in these areas and wish to collaborate with us, please get in touch with me.

More specifically, we are interested in numerical software that runs at the limits of the underlying hardware, where performance constraints dictate engineering/design decisions. Examples include, but not limited to:

  • Large-scale PDE-constrained inverse problems (e.g. Full-Waveform Inversion, commonly used in Geophysical Exploration)
  • Large-scale Deep and Wide Neural Networks (e.g. 3D Image Segmentation, Large-scale data compression, Scientific Machine Learning)
  • Inference or (especially) training of deep learning networks on Edge hardware including mobile phones
  • Combinations of the above, e.g. Physics-Informed Neural Networks

We look at these problems from the perspective of Domain-Specific Languages (DSLs) and their associated compilers, including topics like Automatic Differentiation. We work in diverse application areas including Geophysical exploration as well as Healthcare.

If you would like to do a PhD with me…

I am always on the lookout for motivated people to work with either as PhD students or PostDocs. Please get in touch if you are interested in working with me. There are a few options to fund your PhD at the University of Liverpool:

Other funding opportunities might arise from time to time so please get in touch to ask. You may also get in touch if you have your own means of supporting your study.

If you would like to do a PostDoc with me…

Email me. Now.

If you would like to do a Masters/Undergraduate project with me…

Here are some topics that I’m currently interested in. If any of them sound interesting, email me asking for more information. If you’re super-motivated, you can also think of your own project aligned with these areas but don’t worry if you can’t.

Topics:

  • Fourier Neural Operators for solving PDEs
  • Physics-Informed Neural Networks for solving PDE-constrained inverse problems
  • Segmentation of MRI images
  • Modelling EEG data using deep learning (to be eventually used for the detection of Epilepsy)
  • Compression of scientific data using deep learning