Job opportunities
Centres Inria associés
Type de contrat
Contexte
<p><a href="https://deepinv.github.io/deepinv/">DeepInverse</a> is the open-source PyTorch-based library for solving imaging inverse problems with deep learning. The library implements the entire image reconstruction framework, including efficient forward operators, defining and solving variational problems and designing and training advanced neural networks, for a wide set of domains (medical imaging, astronomical imaging, remote sensing, computational photography, compressed sensing and more). DeepInverse is part of the <a href="https://landscape.pytorch.org/">official PyTorch ecosystem</a>, and is currently used by thousands of imaging scientists and engineers across the world.</p>
<p><em>Documentation:</em> <a href="https://deepinv.github.io/deepinv/">https://deepinv.github.io/</a></p>
<p><em>GitHub repository: </em><a href="https://github.com/deepinv/deepinv">https://github.com/deepinv/deepinv</a></p>
<p>The candidate will work at the IAN team of the Physics Laboratory of <a href="https://www.ens-lyon.fr/">ENS Lyon</a>, under the supervision of <a href="https://tachella.github.io/">Julián Tachella</a> (CNRS), and will work closely with all DeepInverse maintainers (<a href="https://samuro95.github.io/">Samuel Hurault</a>, <a href="https://andrewwango.github.io/">Andrew Wang</a>, <a href="https://github.com/mh-nguyen712">Minh Hai Nguyen</a>, <a href="https://jeremyscanvic.com/">Jérémy Scanvic</a> and <a href="https://www.linkedin.com/in/thibaut-modrzyk-697668221/?originalSubdomain=fr">Thibaut Modrzyk</a>) and <a href="https://github.com/deepinv/deepinv/graphs/contributors">contributors</a>. This position is part of the <a href="https://p16.inria.fr/fr/">P16 program</a>—"a sovereign library ecosystem for AI." led by Inria. The candidate will also receive guidance from Pascal Carrivain (software engineer at INRIA Lyon, <a href="https://team.inria.fr/ockham/fr/">OCKHAM team</a>).</p>
<p><strong>Place</strong></p>
<p>The candidate will be affiliated with the Experimentation and Development Service of INRIA Lyon and hosted at ENS Lyon (46 allée de l’Italie, Lyon, France).</p>
<p><em>Documentation:</em> <a href="https://deepinv.github.io/deepinv/">https://deepinv.github.io/</a></p>
<p><em>GitHub repository: </em><a href="https://github.com/deepinv/deepinv">https://github.com/deepinv/deepinv</a></p>
<p>The candidate will work at the IAN team of the Physics Laboratory of <a href="https://www.ens-lyon.fr/">ENS Lyon</a>, under the supervision of <a href="https://tachella.github.io/">Julián Tachella</a> (CNRS), and will work closely with all DeepInverse maintainers (<a href="https://samuro95.github.io/">Samuel Hurault</a>, <a href="https://andrewwango.github.io/">Andrew Wang</a>, <a href="https://github.com/mh-nguyen712">Minh Hai Nguyen</a>, <a href="https://jeremyscanvic.com/">Jérémy Scanvic</a> and <a href="https://www.linkedin.com/in/thibaut-modrzyk-697668221/?originalSubdomain=fr">Thibaut Modrzyk</a>) and <a href="https://github.com/deepinv/deepinv/graphs/contributors">contributors</a>. This position is part of the <a href="https://p16.inria.fr/fr/">P16 program</a>—"a sovereign library ecosystem for AI." led by Inria. The candidate will also receive guidance from Pascal Carrivain (software engineer at INRIA Lyon, <a href="https://team.inria.fr/ockham/fr/">OCKHAM team</a>).</p>
<p><strong>Place</strong></p>
<p>The candidate will be affiliated with the Experimentation and Development Service of INRIA Lyon and hosted at ENS Lyon (46 allée de l’Italie, Lyon, France).</p>
Mission confié
<p>Deep learning for imaging is revolutionising science, healthcare and engineering, for example by accelerating medical imaging. By contributing to DeepInverse, you will be building and working with cutting-edge tools and algorithms for bringing AI to real-world applications, including diffusion models, foundation models, advanced imaging devices etc. Every day, you will apply skills in engineering, applied math and modern open-source Python programming.</p>
<p>You will also join the vibrant open-source DeepInverse community and participate in France-based and international events such as hackathons and workshops, and will be invited to give tutorials at summer schools, workshops and other events. You will work in close collaboration with an international team of maintainers. Moreover, you will have access to state-of-the-art GPUs via the Inria ABACA and Jean Zay servers, to train or inference state-of-the-art deep learning models for image reconstruction. As part of ENS Lyon, you will access weekly AI seminars and other regular workshops. Finally, you will also have the opportunity to contribute to scientific publications showcasing use cases of the library.</p>
<p> </p>
<p>You will also join the vibrant open-source DeepInverse community and participate in France-based and international events such as hackathons and workshops, and will be invited to give tutorials at summer schools, workshops and other events. You will work in close collaboration with an international team of maintainers. Moreover, you will have access to state-of-the-art GPUs via the Inria ABACA and Jean Zay servers, to train or inference state-of-the-art deep learning models for image reconstruction. As part of ENS Lyon, you will access weekly AI seminars and other regular workshops. Finally, you will also have the opportunity to contribute to scientific publications showcasing use cases of the library.</p>
<p> </p>
Principales activités
<p><strong>Main Activity (approximately 80% of time)</strong></p>
<ul>
<li>Contribute new reconstruction algorithms, imaging operators and benchmarks.</li>
<li>Review pull-requests and manage issues.</li>
<li>Write documentation and new examples for new algorithms, use-cases or datasets.</li>
<li>Improve and maintain the continuous integration.</li>
<li>Train reconstruction models.</li>
<li>Participate in weekly meetings with maintainers.</li>
</ul>
<p><strong>Other Activities (approximately 20% of time):</strong></p>
<ul>
<li>Participate in the organization of hackathons.</li>
<li>Participate in tutorials in summer schools, workshops and other hackathons</li>
<li>Lead training programs for industrial partners</li>
<li>Develop targeted solutions for non-academic partners</li>
<li>Integrate with other open-source libraries (e.g. LazyLinops, etc)</li>
<li>Create training material (videos, blog posts, etc.)</li>
<li>Represent the library within the P16 program</li>
</ul>
<ul>
<li>Contribute new reconstruction algorithms, imaging operators and benchmarks.</li>
<li>Review pull-requests and manage issues.</li>
<li>Write documentation and new examples for new algorithms, use-cases or datasets.</li>
<li>Improve and maintain the continuous integration.</li>
<li>Train reconstruction models.</li>
<li>Participate in weekly meetings with maintainers.</li>
</ul>
<p><strong>Other Activities (approximately 20% of time):</strong></p>
<ul>
<li>Participate in the organization of hackathons.</li>
<li>Participate in tutorials in summer schools, workshops and other hackathons</li>
<li>Lead training programs for industrial partners</li>
<li>Develop targeted solutions for non-academic partners</li>
<li>Integrate with other open-source libraries (e.g. LazyLinops, etc)</li>
<li>Create training material (videos, blog posts, etc.)</li>
<li>Represent the library within the P16 program</li>
</ul>
Compétences
<p><strong>Mandatory skills:</strong></p>
<ul>
<li>Expertise on imaging inverse problems, computer vision or related fields.</li>
<li>Native proficiency in Python and PyTorch, especially applied to computer vision.</li>
<li>Working with and engineering scientific datasets</li>
<li>Strong expertise on modern open-source development (GitHub), including continuous integration (pytest, doctest, docker), collaborative version control, and writing high quality documentation</li>
<li>Be highly independent and proactive, with excellent communication skills</li>
<li>Be collaborative and work well in an asynchronous international team</li>
<li>Having working proficiency in English</li>
</ul>
<p><strong>Desired (but not mandatory) skills:</strong></p>
<ul>
<li>Having contributed to DeepInverse in the past</li>
<li>Experience with CI management tools (GitHub Actions, Sphinx and Sphinx-Gallery).</li>
<li>Experience with CUDA and/or other parallel computing frameworks</li>
<li>Having conducted research on imaging inverse problems and/or deep learning applied to inverse problems</li>
<li>Be proficient in French</li>
</ul>
<ul>
<li>Expertise on imaging inverse problems, computer vision or related fields.</li>
<li>Native proficiency in Python and PyTorch, especially applied to computer vision.</li>
<li>Working with and engineering scientific datasets</li>
<li>Strong expertise on modern open-source development (GitHub), including continuous integration (pytest, doctest, docker), collaborative version control, and writing high quality documentation</li>
<li>Be highly independent and proactive, with excellent communication skills</li>
<li>Be collaborative and work well in an asynchronous international team</li>
<li>Having working proficiency in English</li>
</ul>
<p><strong>Desired (but not mandatory) skills:</strong></p>
<ul>
<li>Having contributed to DeepInverse in the past</li>
<li>Experience with CI management tools (GitHub Actions, Sphinx and Sphinx-Gallery).</li>
<li>Experience with CUDA and/or other parallel computing frameworks</li>
<li>Having conducted research on imaging inverse problems and/or deep learning applied to inverse problems</li>
<li>Be proficient in French</li>
</ul>
Référence
2026-09919
Domaine d'activité