Internship - Image-assisted contrast recovery from highly accelerated post-injection low field MRI

<p>Duration : 6 months</p><p>March 2026 (Flexible)</p><p></p><h4><strong>Skills &amp; Experience</strong><br></h4><p>Training/Profile: Medical Imaging / Deep Learning / MRI Undersampling / Contrast Enhancement</p><p>Expected skills are python/coding, deep learning and signal processing</p><p></p><h4><strong>Key Responsibilities</strong></h4><p>Producing contrast-enhanced MR images typically requires the injection of a gadolinium-based agent, followed by the acquisition of a fully sampled, high-quality volume. However, that acquisition is most often redundant with other images acquired during the same exam, and sometimes even paired with a non-contrast version of the same sequence. Deep learning and MRI reconstruction theory have opened opportunities to generate synthetic contrast-enhanced images by leveraging non-contrast data together with highly accelerated, noisy contrast acquisitions. Such approaches promise substantial reductions in scan time while preserving contrast dynamics critical for clinical interpretation.</p><p>This internship will focus on developing and validating both (1) acceleration/undersampling strategies paired with (2) deep learning reconstruction to recover high-quality contrast-enhanced MRI from a non-contrast image and a strongly accelerated, low SNR contrast-enhanced acquisition.</p><p>- Literature Review: Review contemporary literature in contrast synthesis, image-to-image translation, multi-contrast reconstruction, and accelerated MRI, with emphasis on methods combining multi-contrast inputs.</p><p>- Development and Integration: Design and implement deep learning models to recover diagnostic-grade contrast-enhanced images. Explore strategies to jointly optimize k-space acceleration and reconstruction quality.</p><p>- Data Processing and Evaluation: Develop pipelines for data harmonization, pre-processing, quantitative evaluation, and validation against reference fully sampled contrast-enhanced datasets.</p><p>- Collaboration: Work closely with medical imaging researchers and MRI physicists to refine model assumptions, evaluate reconstruction performance, and integrate findings into broader efforts on accelerated contrast-enhanced MRI.</p><p></p><h4><strong>Who We Are</strong></h4><p>Chipiron is revolutionizing the medical imaging industry by developing light and easily deployable MRI machines powered by an ultra-sensitive quantum detection system. Our mission is to make MRI accessible to everyone, everywhere. We're a team of engineers based in the heart of Paris, and we have only one obsession: building things people want.</p><p></p><h4><strong>Our Culture</strong></h4><ul><li><p>Collaborative Decision-Making;</p></li><li><p>High Standards and Autonomy: We maintain high expectations for our team members and offer complete autonomy, coupled with incentives tied to the company's success, fostering a culture of ownership and accountability;</p></li><li><p>Value of Humility: We prioritize seeking the truth over being right;</p></li><li><p>English-First International Team.</p></li></ul>