Quentin Delfosse

Hi there, my name is Quentin, I am PhD student at the Artificial Intelligence and Machine Learning Lab of TU Darmstadt. I would like to share here my projects, knowledge and interest with people interested in informatics, computer science, machine learning and robotics.

My current research is focused on Neural Networks Architectures, Reinforcement Learning and Robotics. I am aslo a 3D printing enthusiast. Please find more details below, along with other projects that I have done.

Artificial Intelligence

Reinforcement Learning
Deep Learning
Explainable Learning
Data Preprocessing


Computer Vision


3D modeling
3D printing

Other Interests

Game Design
New Technology



2014 - Graduation

I graduated with high distinctions. I passed the french Scientific “baccalauréat”, with high scores in math, sciences, and foreign languages.

2014-2016 Higher School Preparatory Classes

Two years of Higher School Preparatory Classes (intense engineering-based studies).

2016-2019 ENSIMAG Higher School + MoSIG AIW

Graduated a international double diploma ENSIMAG + UGA, specialized in Data Science (Artificial Intelligence and the Web).

2018 - ERASMUS in Germany

ERASMUS Semester Exchange in Germany:
Artificial Intelligence and Robotic Courses (48 ECTS Semester)

1st Research Paper in human prosthetic legs.


2019 - Master Thesis at IAS Darmstadt

Master Thesis on Intrinsic Motivation Models for Robotic Grasping conducted at IAS TU Darmstadt, and publication to IROS workshop.

Since 2020 - PhD on Explainable RL

I am conducting a PhD on Explainable Reinforcement Learning in AIML and IAS labs of TU Darmstadt.



Unsupervised Object Detection

with Optical Flow and Time Consistency

Unsupervised Object Detection

with Optical Flow and Time Consistency

Unsupervised Object Detection

with Optical Flow and Time Consistency

Research on Data Fusion 📈

at Qarnot Computing Paris

Human Locomotion 🦿

Research on Power and Energy needed for locomotion

Grasping learning 🦾

Research on robot grasping using Intrinsic Motivation.
Knowledge transfer from a simulation to Thor, a 3D printed robot.