I'm a very communicative scientist who believes that this is a relevant virtue not only for research teams but also for development teams, as communication, for teamwork, is essential. However, I consider curiosity to be my fundamental characteristic and, like a good scientist, the effort within that curiosity 🔭🌱.
I am passionate about simulating natural phenomena from my small laptop — it's like having a small part of the universe inside my room.
My latest project focused on the development of an open-source scientific package for the fast and accurate simulation of the time-independent wavefunction of a Quantum Harmonic Oscillator, a model widely used in Photonic Quantum Computing. This package, called Fast Wave, was developed as part of my master's research, which is now completed.
I am the co-founder of an indie game and AI studio called SeaFortress, part of the PEPITE Normandie Incubator.
We develop indie games with a strong focus on integrating Artificial Intelligence into gameplay and mechanics. Our games are published on Steam, and we have participated twice in international indie game competitions.
We're currently developing Plinio Needs a Hand, the first game to use Unity Sentis, Unity’s AI library for Computer Vision. This game was showcased at:
I handle all AI-related aspects (Machine Learning and Deep Learning). Our philosophy is: "Strong Games with Smooth Gameplays", and we believe that work should adapt to life, not the other way around.
My journey in Artificial Intelligence and Data Science blends research, applied development, and production-ready solutions.
I specialize in:
- ✅ Risk Modeling and Validation, including Credit Risk, Operational Risk, Market Risk, Microcredit, and Climate & Environmental Risk.
- ✅ Development of challenger models using techniques like MLP, XGBoost, LightGBM, XGBSE, and clustering algorithms such as KMeans, PAM, CLARA.
- ✅ Implementation of probabilistic models like Value at Risk (VaR) and RAROC.
- ✅ Anomaly detection systems, particularly for event logs using machine learning algorithms like Isolation Forest.
- ✅ Advanced Natural Language Processing (NLP) applied to trajectory prediction with Transformer models (Huggingface), treating sequences of geospatial points as sentences.
- ✅ Model Governance, Reproducibility, EDA, Benchmarking, Feature Engineering, and Model Performance Evaluation.
- ✅ Proficient in Python, R, SQL, and distributed systems with tools like RabbitMQ and ElasticSearch.
- Smart Farming with Deep Neural Networks: DOI
- Trajectory Modeling via NLP: DOI — Teach GitHub
- Deep Learning for Trajectory Classification: DOI
- Neuroevolution for Game AI: DOI
This experience combines academic rigor with real-world AI, delivering solutions in Machine Learning, NLP, Risk Modeling, Deep Learning, Data Science, and Model Governance.