MACHINE LEARNING/PRODUCT ENGINEER • Intern • Amsterdam
• Launched an internal initiative for Machine Learning usage implementing semantic search using SBert reducing the tedious workload of finance operations by 70%. • Worked on a POC using LlangChain and Mistral 7*8B via Hugging face using RAG in order to facilitate access to internal information. • Built a succesfull internal app using React and C# significantly improving workflow management for the company. • Participated as a Project Engineer in a project automating 50% of the registration flow in France.
CO-FOUNDER • Val d'Isère
• Co-founded startup implementing predictive maintenance for small companies using AI, overseeing a team of 9, increased productivity of users by 20% on average. • Implemented a Flutter Application and the backend using Firebase responsible for Software Architecture with limited developper resources. • Added Computer Vision algorithms • Was responsible for the design and oversight of the product features.
LEAD DEVELOPER/ PRODUCT OWNER • Freelance • Biarritz
• Led a team of 5 to develop a service for managing coworking space reservations through a WooCommerce site with 100 % of uptime, and more than 300 000€ in revenue over the last 3 years. • Implemented a Flutter with Firebase backend for QR code scanning with more than 1000 daily scans.
COMPUTER VISION ENGINEER • Freelance • Bayonne
• Developed a highly accurate golf ball tracking Computer Vision system and a trajectory prediction algorithm for the Big Tilt platform.Developed a highly accurate golf ball tracking Computer Vision system and a trajectory prediction algorithm for the Big Tilt platform.
Korean Advanced Institute of Science and Technology (KAIST), South Korea
Participated in leading-edge projects at KAIST, widely regarded as the top engineering school in South Korea, known for its research-oriented focus.
Technical University Berlin, Germany
Enrolled in courses on Brain-Computer Interfacing and Machine Learning, taught by renowned experts like Klaus-Robert Müller, one of the pioneers in the field.
ENSEIRB-MATMECA - Graduate School in Electronics, Computer Science, Telecommunications, Mathematics and Mechanics, Bordeaux
Mastered the necessary theoretical and practical aspects of computer sciences. Joined the PhD track, which prepares engineering students for scientific research and its applications.
La Prepa des INP Bordeaux, France
Two year high-intensity study of Mathematics and Physics to prepare for the entry into french engineering schools.
PARTICIPANT
Selected to participate in a Hugging Face and Mistral Hackathon, where an innovative solution using large language models (LLMs) was developed. This solution leverages both vision and language models (VLMs and LLMs) to predict and respond to user needs by analyzing their behaviors. The aim is to use this to enable better collaboration with robots.
DEVELOPER
Currently implementing a Transformer based LLM using Python/Pytorch, the next step is doing the same with the Mamba architecture that doesn't require any Attention Layer while achieving similar performance.
RESEARCHER
Design and Implementation of an RL environment for modelling Inter-Solar system communication using DeepMind Melting pot, RLlib, and self-playing IPPO agents, trying to answer the eternal question of : "Where are the Aliens"
RESEARCHER
Developed and implemented a novel dimensionality reduction technique using Graph Embeddings to preserve data substructures effectively. The algorithm performed on par or better than the SOTA while being extremely simple.