Internship – Engineer in Fair and Secure Federated Learning (M/F) – 6 months

Build a future of trust together

Thales is a global leader in high technologies specializing in three sectors: Defense & Security, Aeronautics & Space, and Cyber & Digital. It develops products and solutions that contribute to a safer, more environmentally friendly, and more inclusive world. The Group invests nearly 4 billion euros per year in Research & Development, particularly in key innovation areas such as AI, cybersecurity, quantum, cloud technologies, and 6G. Thales has nearly 81,000 employees in 68 countries.

Our commitments, your benefits

• Our technological expertise
• Our attention to employee work-life balance
• An inclusive and benevolent environment
• Recognized societal and environmental commitment (Thales Solidarity, CAC 40 ESG index…)

Your daily life

Thales Research&Technology, our research center in hardware domains, and CortAIx Labs, our research center in AI and digital, are tasked with proposing breakthrough innovations, maintaining and increasing technological leadership, and ensuring competitiveness for the group.

Collaborative learning brings together machine learning methods that allow multiple data owners to collaborate to train a model without moving data, thus not directly exposing it.

More specifically, federated learning is a collaborative learning method during which owners update a model on their data, with each data owner's updates then being aggregated to update a global model, which is then retrieved by the data owners to perform a new iteration of model updating on local data, until the global model converges.

Collaboration can occur during learning, which is referred to as federated learning.

In this context, one of your objectives will be to focus on the case of a malicious participant who wishes to act as a "FreeRider," meaning not participating in the local model improvement and only returning noise. Your mission will therefore be to explore existing work aimed at detecting this type of behavior.

A second part of your internship will cover another risk: that of poisoning data and then creating a backdoor. Data poisoning aims to degrade model performance, and a backdoor aims to create a flaw known only to the attacker in the model's behavior.

Your profile

INTERNSHIP – Engineer in Fair and Secure Federated Learning (M/F) – 6 months

Do you want to participate in a research project on topics at the intersection of AI and Cyber?

Does the idea of being integrated into a team composed of data science and cybersecurity researchers appeal to you?

Student in M2, Engineering School, or equivalent training, you have knowledge in:

• Cybersecurity;
• Generative AI;
• Electronics;
• English (oral and written – B2/C1 expected).

Rigor, creativity, and a synthetic mind are assets you are recognized for?

Then this internship is for you!

All our internships are conventional and subject to a gratification whose amount is determined according to your level of study.

Thales, a Handi-Engaged company, recognizes all talents. Diversity is our best asset. Apply and join us!