About me - Merry Chrismas🎅
From research to real‑world impact — LLMs, time‑series foundation models, and trustworthy AI for high‑stakes applications.
Industrial PhD candidate in computer science and finance at the University of Luxembourg, working on machine learning and large language models with real-world applications in finance and insurance.
AI enthusiast, tech geek, and hobbyist pilot-in-training, interested in bridging rigorous research with practical, high-impact deployments.
BIO - My Name On Remote Sensing Images On Earth 👇
- BIO - My Name On Remote Sensing Images On Earth 👇
- Publications snapshot
- Publications timeline
- Industry and collaboration
- Projects and skills
- Academic path
- Beyond research
I am an engieer focusing on fine‑tuning and other applications with machine learning (ML) and large language models (LLMs), in collaboration with Foyer Group, a capital investment and insurance company. My broader background spans deep learning, mathematics, statistics, complex systems modeling, and optimization, with training from Paris‑Saclay University, Université Toulouse III – Paul Sabatier, and ISAE‑SUPAERO.
My main research interests revolve around vision‑language models (VLMs) and large language models (LLMs), which are typically Transformer‑based architectures used in real‑world settings, including text classification, interpretability, machine translation, LLM pre‑training, and post‑training (alignment and fine‑tuning).
Publications snapshot
Publications timeline
Industry and collaboration
Alongside the PhD, there is ongoing collaboration with Foyer Group as a Talent Research Partner, bringing LLM and ML research into production‑grade insurance and financial workflows.
Previous industry roles include AI & Data R&D scientist in Paris, cloud architecture and full‑stack development at Kingsoft Cloud, and software development engineering at Aviage Systems.
These experiences span Python and PyTorch for deep learning, distributed and cloud systems, as well as high‑performance implementations in C/C++ with CUDA and OpenMP.
They also inform interests in DevOps and MLOps, covering Docker, GitLab, Linux, and practical LLM fine‑tuning pipelines.
Projects and skills
Selected projects include medical multimodal image fusion using low‑rank representation and CNNs, Grad‑CAM–based interpretability for MRI classification, and multi‑threaded solvers for large‑scale linear systems.
These projects emphasize both methodological novelty and rigorous empirical evaluation on real datasets.
Technical skills cover Python (PyTorch, Django), Java, C/C++ (CUDA, OpenMP), Matlab, R, SQL, HTML/CSS/JavaScript, Dart, LaTeX, Git, Flutter, and Vue.js, with certifications in C programming, SQL databases, neural networks, GANs, and NLP.
Current interests also include reinforcement learning for reasoning‑oriented LLMs and MLOps for reliable deployment of advanced models.
Academic path
Before the PhD, the academic journey combined several advanced master’s programs in France.
These include the M.Sc. in Mathematics, Vision, Apprentissage (MVA) at Université Paris‑Saclay, a M.Sc. in Statistics and Mathematics at Université Toulouse III – Paul Sabatier, and a M.Sc. in Complex System Modeling and Simulation at ISAE‑SUPAERO.
This multidisciplinary training covered topics such as asymptotic statistics, machine learning theory, probabilistic graphical models, signal processing, optimization, inverse problems, and multi‑physics simulation.
It provides the mathematical and computational foundation for current research in LLMs, reasoning models, and large‑scale experimentation.
Beyond research
Outside of research, personal interests include aviation and flight training as a hobbyist pilot‑in‑training, as well as skiing and endurance events.
Activities such as completing the Paris Schneider Electric Marathon and prior student leadership reflect long‑term commitment and discipline beyond the lab.
