Supervisor: Matthieu Simeoni (LCAV)
Keywords: scientific computing, computational imaging, GPU computing,
inverse problems, convex optimisation
Description: Full-time summer student assistantship (July 1st to August 31st) as a Python developer. The student (which must be currently studying at EPFL) will be expected to help with the development of Pycsou, a Python 3 package for solving large scale linear inverse problems. In particular, the student will be responsible of:
- Introducing support for Cupy and Dask arrays, allowing for GPU and out-of-core, distributed computations,
- Making Pycsou interoperable with the deep learning library PyTorch.
- Extending the package API by implementing new linear operators (in particular wavelet transforms), cost functionals and algorithms.
- Improving/maintaining the package documentation.
- Developing new tutorials in the form of Jupyter notebooks.
Desired Skills: Numerical and Scientific computing with Python 3 on CPU/GPU. Good knowledge of the Python scientific packages (Numpy, Scipy) and possibly their clones for GPU or out-of-core/distributed computing (Cupy, Dask). Some knowledge of convex optimisation and inverse problems is a plus.
Salary: Fr. 24.00/ h (flexible number of hours)
Levels: Master (, possibly Bachelor if sufficient experience with Python)
Sections: All
Deliverables: peer-reviewed Python code, documentation, Jupyter notebooks.
Type of Work: programming, package maintenance.
Location: Flexible (EPFL campus or remote)