Engineer - Scientific programmer in privacy-preserving federate learning with applications in o[...]

European Commission; France; Sur place; EUR 30 000 - 40 000; Plein temps

Résumé du poste

Une institution de recherche en France recrute un chercheur en informatique pour travailler sur des algorithmes de protection de la vie privée. Les responsabilités incluent le développement d'algorithmes et leur application à des cas médicaux, dans le cadre d'une équipe internationale. Le candidat idéal aura une solide formation en informatique et sera intéressé par la recherche en intelligence artificielle. Des avantages tels que plusieurs semaines de congés annuels et la flexibilité de travail sont offerts.

Prestations

Remboursement partiel des frais de transport

7 semaines de congés annuels + 10 RTT

Télétravail et horaires flexibles

Accès à des formations professionnelles

Couverture sociale

Qualifications

• Forte formation en informatique et en mathématiques appliquées.
• Intérêt pour la recherche, notamment sur les questions de confidentialité.
• Expérience dans le codage mathématique efficace et les modules de sécurité souhaitée.

Responsabilités

• Développer des algorithmes et modules pour l'apprentissage machine préservant la confidentialité.
• Tester et appliquer les algorithmes à des cas d'utilisation médicale.
• Collaborer avec des partenaires pour le développement de logiciels open-source.

Connaissances

Développement d'algorithmes

Cryptographie

Statistiques

C/C++

Python

Analyse algorithmique

Formation

Diplôme en informatique

Description du poste

Inria, the French national research institute for the digital sciences

Organisation/Company: Inria, the French national research institute for the digital sciences

Research Field: Computer science

Researcher Profile: First Stage Researcher (R1)

Country: France

Application Deadline: 11 Jul 2025 - 00:00 (UTC)

Type of Contract: Temporary

Job Status: Full-time

Hours Per Week: 38.5

Offer Starting Date: 1 Jul 2025

Funding: Not funded through the EU Research Framework Programme

Reference Number: 2024-07159

Research Infrastructure: No

Job Description

While AI techniques are becoming more powerful, there are growing concerns about potential risks and abuses. This has led to increased research in areas such as privacy-preserving machine learning, explainable AI, fairness, and data protection legislation.

Privacy-preserving machine learning focuses on developing models from data without revealing the data itself. Concepts like local differential privacy help limit information disclosure.

The MAGNET team is involved in projects like TRUMPET, FLUTE, and REDEEM, aiming to develop algorithms for secure, privacy-preserving federated learning, especially in scenarios with potentially malicious participants. TRUMPET and FLUTE target oncology applications, while REDEEM has a broader scope.

The recruited engineer will collaborate with team members and international partners, contributing to the development of platforms and open-source libraries by designing architectures and modules related to privacy-enhancing technologies (PETs) and privacy assessment tools, based on MAGNET's scientific advances.

All software developed will be open-source by default.

Tasks

• Develop algorithms, including cryptographic or statistical modules, and modules supporting knowledge discovery pipelines.
• Test algorithms through benchmarking and experimentation.
• Apply algorithms to medical use cases, such as lung, head & neck, or prostate cancer.
• Research new algorithms for data privacy reasoning.
• Analyze and transform algorithms and queries automatically.
• Design and prototype key algorithms.
• Integrate implementations into the FLUTE platform.
• Work with distributed algorithms, software design, and development (Python and/or C/C++).
• Understand process models and probabilistic reasoning techniques.
• Comprehend programming language internals, such as abstract syntax trees.

Languages

French: Basic

English: Good

Additional Benefits

• Partial reimbursement of transport costs
• 7 weeks annual leave + 10 RTT days + options for exceptional leave
• Teleworking and flexible hours
• Provision of professional equipment
• Participation in social, cultural, and sports activities
• Access to vocational training
• Social security coverage

Candidate Profile & Selection

Seeking candidates with a strong computer science background, interested in research, and capable of handling mathematical aspects of privacy. Experience with efficient mathematical coding, security modules, AI libraries, or algorithm analysis is advantageous.

Application requires CV, cover letter, recommendation letters, and publication list.