Skip to content
View Titouaaaan's full-sized avatar

Highlights

  • Pro

Block or report Titouaaaan

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
Titouaaaan/README.md

Hey, I'm Titouan 👋

M2 student in AI & Machine Learning at Sorbonne Université (Paris),with a strong focus on deep learning. I love the mathematical side of ML, the kind of stuff that makes models actually make sense. I try to open source everything I build!

My main interest right now is reinforcement learning but i've also worked a lot on computer vision and large language models. I also have a soft spot for applying all of this to video games when possible.

31/03/2026: I am currently working on world models and physics informed networks to be used in Model Based RL in the context of my M2 internship (most work is kept private at the moment).


Research Experience

  • AI Research Assistant @ AI Robolab, 2024 — Built a multi-agent LLM system for low-resource language learning (Luxembourgish). Co-authored a peer-reviewed paper presented at PRIMA 2024 in Kyoto, Japan.
  • AI Research Intern @ ONERA / ISIR (Sorbonne), 2026 — Investigating model-based deep reinforcement learning (MBRL) for drone control, with a focus on integrating physics-based priors into hybrid models. Working on algorithms like TD-MPC and PETS applied to fixed-wing UAV systems in simulation.

🚀 Projects

Project Description
RL-Souls Model-free RL algorithms trained to beat Dark Souls III using a custom Gymnasium environment
NN from Scratch Fully modular neural network library built with NumPy only: implemented conv layers, pooling, SGD, PCA, K-Means
LLM-RAG Experimental RAG pipeline combining SPLADE retrieval, BM25, and Chain-of-Thought reasoning
GRLA Reproduction of a lightweight image super-resolution architecture bridging softmax and linear attention via Gaussian RBF kernels (ICLR 2025 submission)
LLM Multi-Agent System LangGraph-powered multi-agent system, backend code for an accepted PRIMA conference paper
Portfolio App Self-hosted fullstack app built with Next.js and FastAPI

🛠️ Tech Stack


📚 Education

MSc AI & Machine Learning — Sorbonne Université (2023 – Present) (I took a gap year in 2024!) Probabilistic Modeling · Deep Learning · NLP · Reinforcement Learning · LLMs

BSc Computer Science — University of Luxembourg (2020 – 2023)
Competitive Programming Club · Linear Algebra · Algorithms & Complexity


My Awesome Stats

Pinned Loading

  1. RL-Souls RL-Souls Public

    Using Reinforcement Learning to play Dark Souls III

    Python 1

  2. ELL-MMA ELL-MMA Public

    Beyond Chatbots: Enhancing Language Learning Through Multi-Agent Systems and LLM

    Python 3

  3. DIY-Neural-Network DIY-Neural-Network Public

    Implementation of a neural network from scratch

    Jupyter Notebook 3

  4. M2-MIND M2-MIND Public

    contenu des cours + projets du M2 DAC/MIND

    Jupyter Notebook 1

  5. redscarfx/GRLA redscarfx/GRLA Public

    Reproduction of a project about bridging softmax and linear attention via gaussian rbf kernel for light weight image super-resolution

    Jupyter Notebook 4

  6. LLM-RAG LLM-RAG Public

    Experimental implementation of a Retrieval-Augmented Generation (RAG) pipeline combining SPLADE-based retrieval and LLMs, with dataset generation, retriever training, and extended experiments

    Python 2 1