Join us as an Intern!

CEA Tech Corporate from CEA Tech on Vimeo.

 As an intern at CEA, you will have the opportunity to work within a world-renowned research environment. Our teams comprise passionate and dedicated experts, providing a conducive setting for learning and collaboration. You will have access to cutting-edge equipment and top-tier research resources to successfully accomplish your missions.

Job Description:

Within the Signals and Sensor Systems Laboratory, you will work with a team of 10 researchers.

Data-driven learning (machine learning) has revolutionized a wide range of fields. Compared to other knowledge-based algorithms, they have demonstrated their ability to perform complex tasks (object detection, translation, text synthesis). However, training these models requires a significant amount of data, especially for deep learning networks. For instance, ChatGPT was trained on several hundred gigabytes of data.

What can be done for other tasks where training data is difficult or costly to acquire? Or where it is insufficient? In recent years, a new paradigm has emerged, combining data with prior knowledge. Indeed, we often have knowledge about the task that is not contained in the training data. We know, for example, that the model output is always positive, that it must respect physics, etc. Adding this knowledge would allow us to achieve the same performance with less training data or obtain a more performant model. One way to add this knowledge has gained significant attention: physics-informed neural networks (PINNs). This method enables obtaining a solution that respects both the training data and a physical equation.

Your internship will consist of applying this method to the case of a metallic beam deformation. To this end, several training datasets (experimental and synthetic) have already been acquired. The laboratory has also artificially created defects on this beam, which need to be localized.
Your tasks will include:

1. Conducting a literature review on PINNs
2. Understanding and formatting the training datasets
3. Training a classical neural network
4. Training a physics-informed neural network
5. Analyzing your results and potentially drafting a publication
Throughout your internship, particular attention will be paid to the organization of your code, version control, and documentation, as well as the critical analysis of the results obtained for different algorithm parameters.

What do we expect from you?

You are pursuing a Master's degree (2nd year) or a 3rd-year engineering degree in the following fields: computer science, mathematics, telecommunications.

You are passionate about scientific and technological research and are recognized for your rigor, curiosity, and enthusiasm for solving open problems.

You possess knowledge in machine learning and deep learning and ideally have some experience in these fields. Any prior experience in research will be valued, particularly in reading scientific articles.

Join us and come develop your skills and acquire new ones!

Still have doubts? We offer:

The opportunity to work within a world-renowned organization in scientific research,
A unique environment dedicated to ambitious projects addressing current societal challenges,
An experience at the forefront of innovation with significant industrial development potential,
Exceptional experimental means and high-quality supervision,
Real career opportunities after your internship,
A position in the heart of the Grenoble metropolitan area, easily accessible via soft mobility favored by CEA,
An 85% contribution to public transportation costs,
A recognized work-life balance,
A company restaurant,
A diversity and inclusion policy,
In line with CEA's commitments to integrating people with disabilities, this position is open to all. CEA offers arrangements and/or organizational possibilities for the inclusion of workers with disabilities.