Julien Perez - Machine Learning | AI Research Professor

🎓

Julien Perez

Full Professor | Head of AI | Pedagogy Research Program

About Me

Seasoned AI research leader with 15+ years of experience in academia and industry. Currently serving as Full Professor at EPITA and Head of AI & Pedagogy Research Program at IONIS Education Group. Expertise spans reinforcement learning, natural language processing, robotics, and large language models, with a proven track record of translating cutting-edge research into impactful real-world applications.

My research has achieved state-of-the-art performance on major benchmarks and has been recognized with international awards. I hold 17 patents in machine learning and NLP, and have led teams ranging from 6 to 30 researchers across academia and industry. My current focus includes developing AI-driven educational tools, advancing robotic learning with LLMs, and mentoring the next generation of AI researchers.

Recent Publications

AKD: Adversarial Knowledge Distillation for Large Language Models Alignment on Coding Tasks

Adversarial knowledge distillation framework that stress-tests and distills Code-LLMs into smaller, more robust and reliable models.

May 5, 2025

Unsupervised Skill Discovery for Robotic Manipulation through Automatic Task Generation

Unsupervised skill discovery approach using automatically generated tasks and asymmetric self-play to learn composable robotic manipulation behaviors.

Oct 7, 2024

SLIM: Skill Learning with Multiple Critics

Multi-critic actor-critic approach that combines multiple reward functions to discover safe, useful robotic manipulation skills.

Feb 1, 2024

DiffVersify: a Scalable Approach to Differentiable Pattern Mining with Coverage Regularization

Scalable differentiable pattern mining approach with coverage regularization for large-scale datasets.

Jan 1, 2024

Controlling the Solo12 Quadruped Robot with Deep Reinforcement Learning

End-to-end deep reinforcement learning controller for the open-source Solo12 quadruped robot, validated on real hardware.

Aug 2, 2023

LARG, Language-based Automatic Reward and Goal Generation

LLM-based approach that automatically converts text-based task descriptions into reward and goal-generation functions for robotic manipulation.