Fully Funded PhD Scholarship: Multimodal Signals, Université de Lorraine, France

Summary

The MULTISPEECH team at LORIA (Université de Lorraine, Inria, CNRS) is recruiting a fully funded PhD researcher under the AI Grand Est ENACT Research Chair. The project focuses on developing novel privacy-preserving methods for clinical spoken language understanding (SLU) by integrating speech processing, multimodal machine learning, voice anonymization, and speech/audio large language models (Speech-LLMs) to ensure patient data confidentiality while retaining critical diagnostic biomarkers.

Fully Funded PhD Scholarship: Privacy-Preserving Speech Understanding with Multimodal Signals for Clinical Applications

Designation

• PhD Candidate / Doctoral Researcher (Full-Time, Fully Funded)

Overview Table

Research Area

• Spoken Language Understanding (SLU) & Speech Processing: Voice anonymization, spoken-content de-identification, and acoustic feature extraction.
• Multimodal Machine Learning & AI for Healthcare: Speech and audio Large Language Models (Speech-LLMs), integration of clinical signals (e.g., respiratory distress cues from emergency calls, EHR metadata, physiological data).
• Trustworthy & Privacy-Preserving AI: Measuring privacy–utility trade-offs, defensive learning, and robustness against privacy attacks in digital health.

Location

• Laboratory: LORIA (Laboratoire Lorrain de Recherche en Informatique et ses Applications), Inria Nancy – Grand Est, CNRS.
• Address / City: Villers-lès-Nancy (Nancy metropolitan area), Grand Est, France.

Eligibility & Qualifications

• Academic Background: Master’s degree (M.Sc. / M2) or an equivalent Engineering degree in Computer Science, Artificial Intelligence, Speech/Signal Processing, Applied Mathematics, Data Science, or a related discipline.
• Technical Skills:
  • Proficiency in programming with Python.
  • Hands-on experience with deep learning frameworks such as PyTorch and Hugging Face / Transformers.
  • Solid foundations in machine learning, NLP, speech processing, or multimodal modeling.
• Preferred Competencies: Familiarity with voice privacy/anonymization techniques, biomedical data processing, or large audio language models is advantageous.
• Languages: Working proficiency in English (oral and written); French language skills are helpful but not mandatory.

Scholarship Description & Benefits

• Full Financial Coverage: 100% funded by the AI Grand Est ENACT Research Chair.
• Remuneration: Monthly gross salary of €2,300 including social security coverage.
• Research Environment: Access to high-performance computing clusters (HPC), collaborative ties with clinical partners (e.g., SAMU emergency services), and participation in leading international AI and speech conferences (INTERSPEECH, ICASSP, NeurIPS, etc.).

How to Apply

Interested candidates should prepare and submit their application dossier directly via email to the principal supervisor:

1. Email Recipient: natalia.tomashenko@inria.fr
2. Subject Line: [PhD Application - ENACT] Privacy-Preserving Speech Understanding - <Your Full Name>
3. Required Documents:
   • Detailed Curriculum Vitae (CV).
   • Motivation / Cover letter explaining research interests and alignment with the topic.
   • Academic transcripts and diplomas (Bachelor’s and Master’s).
   • Contact details of 2 academic/professional references (or letters of recommendation).
   • Master’s thesis summary or relevant code/publication links (if available).

Last Date for Apply

• Primary Deadline: August 30, 2026 (Applications reviewed on a rolling basis until the position is filled)
• Extended Deadline: September 20, 2026 (Applicable if no suitable candidate is selected in the initial round)

Link