Ph.D. in computer science, I worked for three years on a smart home project in the team ACES at INRIA Rennes (France) and in collaboration with EDF R&D. The goal of the project was to develop context-aware services while preserving users’ privacy. I proposed an application of the belief functions theory over a wireless sensor network for context recognition. Within this framework, I implemented algorithms to compensate for the instability of sensor measures and the data loss due to wireless communications.
During my post-doc at Nimbus Centre in Cork (Ireland), I worked within the European project Tribute. We investigated novel methods for the detection of occupancy in public buildings (offices, libraries, etc.). I proposed a system using Wi-Fi sniffers gathering probe packets sent by Wi-Fi devices to locate those devices and estimate occupancy based on a probabilistic approach. I’ve also used machine learning to count people based on simple PIR sensor binary events.
I worked one year at Xanadu Consultancy Limited as a data scientist. I proposed, designed, and implemented (in Python) a generic user profiling engine relying on micro-services. This engine is used to build data products for Matchbook.com.
After a one year break due to health issues, I restarted academic research at Insight (Centre for Data Analytics). I’m currently part of the European project LOGISTAR, working on predictions for logistics.
My research interests are broad: ubiquitous computing, data fusion, Dempster-Shafer theory, machine learning applications, and still astronomy and astrophysics (though I’m not publishing in those domains).