Hosting institution: LIS Lab, CNRS, Aix-Marseille University
Collaborating institution: LPT Lab, CNRS, Toulouse University
Financial support: CNRS AISSAI Center
Duration: 18 months
Start Date: As soon as possible
Application Deadline: October 15, 2025
Project leaders: Prof. Hachem Kadri (LIS, Aix-Marseille University) and Prof. Ion Nechita (LPT, CNRS)
Deep learning models, and in particular large language models (LLMs), have demonstrated remarkable capabilities but remain limited by their heavy computational requirements, lack of interpretability, and tendency to produce unreliable outputs (“hallucinations”).
This project investigates whether quantum computing can enable a new generation of generative AI models that are more efficient, theoretically grounded, and explainable. The research sits at the crossroads of quantum information science, tensor network theory, and generative modeling (transformers, diffusion models, etc.).
Key research directions include:
- Quantum tensor networks for parameter-efficient fine-tuning of LLMs
- Exploring how collapse phenomena in LLMs relate to entanglement concepts and how these insights can improve network expressivity.
The project is supported by the CNRS center AISSAI (AI for science, science for AI) and is hosted by the computer Science Lab LIS at Aix-Marseille University. It is a collaboration with the physics lab of Toulouse LPT, offering a vibrant interdisciplinary research environment.
The postdoctoral researcher will be based at LIS in Marseille but will collaborate closely with colleagues at LPT in Toulouse. Short research stays in Toulouse will be planned as part of the project to foster collaboration.
Job description
• Conduct cutting-edge research at the intersection of quantum information, machine learning, and generative AI
• Design and implement algorithms for quantum-inspired and quantum-enhanced generative models
• Investigate theoretical foundations of tensor networks, entanglement, and collapse phenomena in deep learning
• Collaborate with an interdisciplinary team of AI, quantum physics, and applied mathematics researchers
• Contribute to open-source tools, datasets, and benchmarks
• Present and publish results in leading conferences and journals
Required Qualifications
• PhD in one of the following areas (or related fields):
• Machine learning / deep learning
• Quantum computing / quantum information
• Applied mathematics, computational physics, or computer science
• Strong programming skills (e.g., Python, PyTorch, TensorFlow)
• Experience in one or more of the following areas:
• Generative AI models (transformers, diffusion models, GANs, etc.)
• Tensor networks or quantum algorithms
• Theoretical or applied aspects of quantum information
Desired Qualifications
• Experience with parameter-efficient fine-tuning (e.g., LoRA, adapters, or related methods)
• Knowledge of quantum machine learning or tensor networks
• A strong track record of publications in high-impact journals or conferences
• Interest in developing open-source tools
Benefits
• Salary and social benefits according to French CNRS regulations
• Opportunities to attend conferences, workshops, and short-term research visits
• Collaborative academic environment within LIS (Marseille) and LPT (Toulouse)
How to Apply
Please send the following materials in a single PDF file to hachem.kadri [at] univ-amu [dot] fr and ion.nechita [at] univ-tlse3 [dot] fr with the subject line "Postdoc Application – Quantum Generative AI":
• A cover letter describing your research experience and motivation for applying
• CV, including a list of publications and relevant projects
• Contact information for at least two references
Applications will be reviewed on a rolling basis until the position is filled.
