Mathéo Trebosc
Toulouse, France
Maths & AI master's student at Toulouse School of Economics. atlasfeed.fr
About me
Master's student in Mathematics and Economic Decision at Toulouse School of Economics, I work at the intersection of econometrics, statistics, and machine learning. My favourite area is Reinforcement Learning.
The projects I take part in are ones I genuinely care about. My GitHub is full of experiments and projects...
I follow research closely in Reinforcement Learning, Embodied AI, and World Models, and I am looking for a 6-month internship starting May 2027.
Work
All projects →Monde (live geopolitical globe)
A 3D globe of conflicts and political events fed by 20 live open sources, with a terrain renderer written from scratch in Rust and a measured pipeline that turns news articles into verifiable event records.
Causal Impact of Plain Cigarette Packaging
Causal effect of plain packaging (2017) on cigarette consumption, healthcare utilisation, and smoking cessation — Master's project at TSE, in R.
Insider-trading scoring (USA)
Retrieval, cleaning, and quantitative scoring of insider transactions declared in the United States.
AI contract analysis (OCR & PII)
Financial data extraction (OCR), PII pseudonymisation, and structural contract analysis | Innovation internship at Doxallia.
Atlas (country data API)
A country-profile API and dashboard that unifies 18 open data sources, with forecasts, citable snapshots and licence-aware sourcing.
Currently reading
- Reinforcement Learning: FoundationsShie Mannor, Yishay Mansour, Aviv TamarSource ↗
- Reinforcement Learning: Theory and AlgorithmsAlekh Agarwal, Kianté Brantley, Sham M. Kakade, Nan Jiang, Wen SunSource ↗
Loading activity…
What I write & read
See all →My Publications
Nathan Stouque, Lara Le Bihan, Mathéo Trebosc
Causal effect of France's 2017 plain packaging on cigarette consumption, healthcare utilisation and smoking cessation, estimated from sales data and health records rather than self-reported surveys. Consumption declines by about 3.7% at the baseline estimate, and by up to about 7.6% by 2019, alongside a sharp rise in nicotine-replacement therapy use and smoking-cessation consultations. Applied Econometrics project (Master 1, Toulouse School of Economics), supervised by Angélique Acquatella.
Reviews
Reinforcement Learning: Foundations
Shie Mannor, Yishay Mansour, Aviv TamarbookReading course · Reinforcement learningReading this book for my Reading Course on world models and model-based RL, supervised by Achraf Azize. Review to come.
Reinforcement Learning: Theory and Algorithms
Alekh Agarwal, Kianté Brantley, Sham M. Kakade, Nan Jiang, Wen SunbookReading course · Reinforcement learningReading this book for my Reading Course on world models and model-based RL, supervised by Achraf Azize. Review to come.
- paperAI · Frontier · Kimi · Attention · LLM · Open-weight · Chinese
Kimi K3 shows that open weights are a good thing. It shuffles the cards once again, and it is not over.