Clement Marie

I work on deep learning for perception, moving toward world models for autonomous driving.

Seeking PhD opportunities
I am looking for a PhD in deep learning, with a focus on perception and world models for autonomous driving. Get in touch.

MVA master's student at ENS Paris-SaclayResearch Engineer Intern at CEA List

Portrait of Clement Marie

From perception to world models

My work spans visual localization, 3D perception, and multimodal representation learning.

I study how deep learning systems can combine semantic and geometric cues across sensors and viewpoints. Building on this work in perception, I am interested in moving toward persistent world models that represent dynamic scenes over time for autonomous driving.

Deep learning3D perceptionVisual localizationMultimodal learningWorld models

Research experience

Saclay, France

Research Engineer Intern

CEA List· LVA

I am building a persistent camera–radar 3D state, including the multimodal architecture and the complete nuScenes training and benchmarking pipeline for 3D detection and semantic segmentation.

Munich, Germany

Research Engineer Intern

Huawei Research Center

I improved position/yaw recall@1m by 4% over the OrienterNet baseline. I extended the model with Fourier encodings and graph neural networks for geometry-grounded, viewpoint-invariant global visual localization and delivered reproducible Docker environments.

OrienterNet repository ↗

Selected projects

01

Personalized HRTF prediction

I won 2nd place and a $3k award at Huawei TechArena by training a multimodal model to predict personalized spatial audio from paired ear images and HRTF data.

GitHub ↗
02

Spike Transformer: Neural Position Decoding

We achieved the best score at the ICM Hackathon 2026 with a transformer that decodes a mouse’s position from neural recordings.

GitHub ↗

Education

École Normale Supérieure Paris-Saclay

Master’s degree · M2 Mathematics, Vision, Learning (MVA)

Shanghai ARWU 2026 · #13 worldwide

The MVA is one of the best AI master’s programs in France and Europe. Led by ENS Paris-Saclay, it connects the mathematical foundations of machine learning with computer vision and data modelling.

I earned 85.7/100 with a 4.0 GPA.

Selected results · 8 courses
  • Deep Learning15/20
  • Geometry Processing & Geometric Deep Learning17.9/20
  • Probabilistic Graphical & Deep Generative Models16.5/20
  • Detection Theory17/20
  • Remote Sensing Data16.5/20
  • Large Language Models for Code and Proof19.7/20
  • Interactions: Complexity, Stochasticity, Extrema & Rare Events19/20
  • Audio Signal Analysis15.5/20

Université Paris-Saclay

Master’s degree · M1 Artificial Intelligence

Shanghai ARWU 2026 · #13 worldwide

I earned 75.1/100.

Selected results · 8 courses
  • Foundational Principles of Machine Learning15.75/20
  • Machine Learning Algorithms15.06/20
  • Deep Learning15.4/20
  • Mathematics for Data Science17.5/20
  • Optimization14.2/20
  • Information Theory15.8/20
  • Research Project16/20
  • Hands-on Natural Language Processing17.2/20

Sorbonne Université

Bachelor’s degree · Mathematics and Computer Science

Shanghai ARWU 2026 · #55 worldwide

I completed an intensive three-year double-major program with 77.5/100.

Selected results · 8 courses
  • Algorithms II15.42/20
  • Algorithms I top 7%16.68/20
  • Introduction to Probabilities15.11/20
  • Data Sciences top 3%19.14/20
  • Numerical Analysis18.4/20
  • Functional Analysis17.1/20
  • Calculability & Decidability18.37/20
  • Lebesgue Integral on Rn top 1.4%19.6/20

English proficiency

TOEFL iBT: 106/120 · CEFR C1