Job Information
Organisation/Company: DIENS
Research Field: Computer science » Informatics; Mathematics » Applied mathematics; Physics » Statistical physics
Researcher Profile: First Stage Researcher (R1)
Positions: Other Positions
Application Deadline: 27 Sep 2026 - 23:59 (Europe/Paris)
Country: France
Type of Contract: Temporary
Job Status: Full-time
Is the job funded through the EU Research Framework Programme? Not funded by a EU programme
Is the Job related to staff position within a Research Infrastructure? No
Offer Description
Job Description
Applications are invited for a full-time, 3 to 12 months pre-doctoral research assistant (RA) position in theoretical machine learning and NeuroAI. The RA will work on developing a new theoretical framework to understand feature learning and parameter dynamics in artificial neural network models of the brain.
This position is funded through a French National Research Agency (ANR) grant.
Project Goal & Research Objectives
The field of machine learning has grown at an unprecedented pace. Neural network models have become indispensable tools across all scientific domains. As reliance on AI continues to increase, there is a growing need to understand how these models learn in order to develop more accurate modelling methods. Yet, the mathematical intractability of neural networks makes this one of the deepest challenges of modern machine learning.
Seminal work has identified two canonical learning regimes: the same architecture may either learn low-dimensional features of their input data (the so-called “rich” learning regime), or random high-dimensional projections (the “lazy” regime). More recent findings suggest that task symmetries and model overparameterisation can be important determinants of the learning regime in which a network operates. Yet, how these factors interact to determine the set of solutions that a neural network can learn remains poorly understood.
In biological modelling applications, such different neural network solutions can be seen as different candidate models of a biological system. Yet, there remains a poor understanding of which solution yields a more faithful model. In particular, in neuroscience applications, it remains unknown which learning regime gives rise to neural networks whose solutions most closely resemble those of biological neurons.
This RA will be targeted towards developing new statistical tools embedded in Riemannian and information geometry to study feature learning in neural network models of biological systems.
Methodology & Key Tasks
The research project will combine theoretical analysis and computational modelling, focusing primarily on:
- Using statistical and Riemannian geometric tools to characterise feature learning
- Using this framework to study feature learning in neural network models used in neuroscience and other biological data applications
Research Environment & Supervision
- Host Institution: The RA will be physically hosted at the Centre Sciences des Données (CSD) at École Normale Supérieure (ENS - PSL), Rue d’Ulm, Paris. The CSD provides a vibrant, world-class theoretical research environment bringing together experts across computer science, physics, and mathematics, with dedicated centre-wide access to high-performance computing resources (Jean-Zay supercomputer).
- Collaboration: The RA will also work in close collaboration with the quantitative biology (QBio) institute at ENS, to expose them to leading theoretical research in neuroscience.
- Supervision: The RA will be supervised by Arthur Pellegrino (PI, AI Fellow at ENS/PR[AI]RIE)
- Mentorship & Mobility: The candidate will benefit from a highly supportive supervision environment, including regular group seminars, opportunities to co-mentor Master's students, dedicated funding to present research at top ML conferences (NeurIPS, ICML, COLT), and access to international collaborative networks.
Candidate Profile
- Degree: Master’s degree (or equivalent) in Theoretical Machine Learning, Mathematics, Physics or a related quantitative field.
- Mathematical prerequisites: Strong foundations in linear algebra, multivariable calculus, and probability theory (familiarity with differential geometry or dynamical systems theory is a strong plus).
- Technical skills: Proficiency in modern deep learning frameworks (PyTorch, JAX).
- Research skills: High motivation to conduct foundational research in AI theory.
Application Details
- Duration: 3-12 months
- Location: Paris, France (ENS - PSL)
- Start Date: Autumn 2026
How to apply
Please contact us via email arthur.pellegrino@ens.fr with the following items:
- CV (including list of publications)
- Academic transcripts (undergraduate and Master's)
- Cover letter (1-page, focused on research interests)
Where to apply
E-mail: arthur.pellegrino@ens.fr
Requirements
Research Field: Computer science » Informatics; Mathematics » Applied mathematics; Physics » Statistical physics
Education Level: Master Degree or equivalent
Work Location(s)
Number of offers available: 2
Company/Institute: Ecole Normale Supérieure, Département d'Informatique
Country: France
State/Province: Paris
City: Paris
Postal Code: 75005
Street: 45 Rue d'Ulm
Contact
City: Paris
Website: https://www.di.ens.fr
Street: 45, Rue d'Ulm
Postal Code: 75005
This is a Preview Listing…
You must sign in to see the full job description, and to apply.
Manage / Upgrade this job to a Full Job Listing.
Find Your Best Opportunity
Tell them AcademicJobs.com sent you!
