Evangelos Papoutsellis

Welcome to my website

I am Evangelos Papoutsellis (or Vaggelis) working at the intersection of applied mathematics, inverse problems, optimisation, tomography, machine learning, and open-source scientific software.

Research: My work focuses on imaging inverse problems, mathematical optimisation, and deep learning for building scalable algorithms that extract reliable information from complex, noisy, and high-dimensional data. I apply these methods to real-world applications in medical imaging, materials science, tomography, and chemical imaging.

Open-source software: I develop open-source research software that turns mathematical methods into practical, reproducible computational tools. This includes work on the Core Imaging Library (CIL), nDTomo, and the Synergistic Image Reconstruction Framework (SIRF), supporting optimisation, reconstruction, simulation, and analysis workflows for imaging and inverse problems.

Opportunities: I am currently open to new opportunities in computational imaging, scientific computing, applied mathematics, machine learning for imaging, and research software engineering, across academia, research institutes, and industry. I am especially interested in roles involving mathematical modelling, large-scale optimisation, algorithm development, tomographic reconstruction, and open-source scientific software.

For more details please see cv (updated May 2026).

news

Jul, 2026	Invited Talk and Organisation: I was delighted to give an invited talk at 15th AIMS Conference, 2026 Athens, Greece. I was also pleased to co-organise the special session New Developments in Open-Source Software for Inverse Problems with Ander Biguri, which brought together researchers and developers working on open-source tools for computational imaging, and inverse problems. All talks are online here.
Apr, 2026	Accepted Paper: Our paper, “Split, Skip and Play: Variance-Reduced ProxSkip for Tomography Reconstruction Is Extremely Fast,” has been accepted to the 2026 IEEE International Conference on Image Processing (ICIP 2026), 13-17 September 2026, Tampere, Finland. The open-source code is available here.
Mar, 2026	New Paper: A Modular Approach to Stochastic Optimisation for Inverse Problems Using the Core Imaging in collaboration with Margaret A. G. Duff, Jakob S. Jørgensen, Sam Porter, Claire Delplancke, Gemma Fardell, Edoardo Pasca and Kris Thielemans.
Mar, 2026	Accepted Talk I gave a talk at the SyneRBI Symposium on AI and Reconstruction for Biomedical Imaging for our recent work Split, Skip and Play: Variance-Reduced ProxSkip for Tomography Reconstruction is Extremely Fast. All talks are online here. I was also invited to participate at the Hackathon on software for machine learning approaches in biomedical imaging working on CIL/RTK and SIRF/DeepInv integrations.
Feb, 2026	New Paper: Split, Skip and Play: Variance-Reduced ProxSkip for Tomography Reconstruction is Extremely Fast in collaboration with Zeljko Kereta, Kostas Papafitsoros.