Physicist turned machine learning engineer, based in Liège, Belgium 🇧🇪
Interested in physics-flavored ML: earth observation, model predictive control, survival analysis, industrial forecasting.
I started out as a physicist (optics/photonics) and moved into machine learning and data engineering. Today I focus on turning industrial and physical data — electricity grids, sensor streams, geospatial data — into pipelines, forecasts, and dashboards that people can actually use.
I'm currently working on data engineering tooling for data scientists, and I'm looking to collaborate on AI-driven multi-party agreement search (Nash equilibrium).
Languages
Machine Learning / Data Science
NLP / LLMs
Geospatial / GIS
Cloud & Data Engineering
Long-term
- Nash equilibrium searcher — can AI search for agreements between multiple parties? Combines ontologies, optimization, and data engineering. Early stage, bibliography phase.
- LLM app for Raspberry Pi accessibility tools — an app that writes flashable code for assistive devices. On hold.
Applied
- Paragliding site searcher — a geospatial tool to surface flyable sites from open geodata.
- AI-aided electricity management — a chatbot that parses a spoken agenda and optimizes energy consumption accordingly. Work in progress.
Learning
- Housing price predictor — a classic regression project used to practice data engineering fundamentals.
- Epidemic simulation & Bayesian inference
- Sudoku grid reader — grid detection and digit recognition with OpenCV.
- Emotion classifier — deep learning on facial expression data.
📫 pro.dimitrimarchand@gmail.com · 📍 Liège, Belgium