STAGE - Ingénieur.e Recherche IA en segmentation médicale (F/H)

Dassault Systèmes has a strategic ambition to revolutionize life sciences with the development of the virtual twin, and consequently to transform medical practice.

Within the Corporate Research & Sciences department, you will join the "Medical Image Segmentation" team, specialized in applying deep learning technologies to medical images. As a key part of the MEDITWIN project, its mission is to develop medical image segmentation methods using deep learning, leveraging our medical data heritage to serve as a foundation for building virtual human body twins.

Different imaging modalities (MRI, CT-scan, ultrasound, etc.) and various anatomical structures exhibit great variability. Associated training datasets are therefore often more restricted and heterogeneous than in other computer vision domains.

A promising avenue is to design universal models, trained on aggregated data from multiple sources covering diverse segmentation tasks. While these models benefit from a larger volume of data and better generalization capabilities, some tasks may conflict during training, to the detriment of overall performance.

This internship will directly address this issue. It will involve exploring how a modular architecture based on a Mixture-of-Experts (MoE) mechanism can best leverage synergies between different segmentation tasks during training on heterogeneous data.

Your Missions

Your missions will include:

• Conducting and presenting a state-of-the-art review on the topics covered.

• Designing and innovating by developing new methods.

• Implementing and evaluating selected solutions.

You will thus have the opportunity to lead a research project from start to finish, enjoying a high degree of autonomy within a dynamic environment driven by knowledge sharing.

Your Qualifications

Currently pursuing a Master 2 / Bac+5 in Engineering School or University, specializing in Artificial Intelligence, Computer Vision, Computer Science, or Applied Mathematics, and looking for an internship between 5 and 6 months.

Desired technical skills:

• Perfect mastery of Python and common practice of PyTorch.

• Good software development practices: mastery of Git and comfort in a Linux environment.

• Scientific approach:

  • Proven ability to read, analyze, and exploit scientific literature (state-of-the-art, research papers).
  • Strong taste for applied research and experimentation.

Desired professional skills:

• Ability to show initiative, propose research directions, and test new hypotheses.

• Ease in presenting work clearly and structured, and desire to contribute to the team.

• Advanced English proficiency (conversational French is a plus).

Assets

• First academic experience or significant project in Computer Vision and Deep Learning for medical applications.

• Practice with standard deep learning development tools (PyTorch Lightning, Hydra, MLflow, etc.).

• A first immersion in the world of research (final year project, internship in a school/university laboratory, or open source contribution).

Joining us also means

Joining a scientific company at the heart of technological innovation, driven by strong growth for over 40 years.

Main advantages and benefits:

• Collaborative and innovative environment

• International collaboration

• Diversity of technologies, products, and solutions

• Commitment to diversity and inclusion