Prototypes I was involved in


SensApp is a cloud platform to collect and query sensor data.

Sébastien Mosser is the main developper of SensApp. I have been contributing a few services to this platform: import/export from/to CSV, export of SRT to dub videos with the values of sensor that have been captured during those videos, factorization of data to provide clients with sorted and minimized data (yet complete) and a resampling service to change the sampling rate of data (e.g. from 1Hz to 5Hz) to provide clients with much lighter (and compressed with loss) data.

I have also been the main advisor of Jonathan Nain (summer 2013), who has provided a Web Socket interface to SensApp, in addition of the existing HTTP/REST interface, in order to support additional usages of the SensApp platform where plain HTTP is too heavy.

SensApp is hosted on GitHub and available as open-source (L-GPL)


CloudML is a modeling language for the provisioning of cloud resources from different cloud vendors and the deployment, monitoring and adaptation of software stacks on these cloud resources.

Together with Nicolas Ferry and Franck Chauvel, I was one of the core developers of the initial version of CloudML, developed in the context of the REMICS project. Now, Nicolas Ferry, Franck Chauvel and Hui Song are the core developers of CloudML. CloudML is hosted on GitHub and available as open-source (L-GPL)


BVR, the Base Variation Resolution language (formerly CVL, the Common Variability Language), is a modeling language to express and enact variability in any existing modeling language, such as the UML or Domain-Specific Modeling Languages. I have used the initial version of CVL in the context of the CESAR project, and was involved in the new version of CVL/BVR.

Anatoly Vasilevskyi and Daisuke Shimbara are the core developers of BVR, while Oystein Haugen is responsible for the definition of the BVR metamodel. BVR is currently incubated on a private GitHub repository, and shall be publicly released as open-source later on.

Aspect-Oriented Modeling weavers

I have been the main contributor of SmartAdapters, an EMF-based aspect model weaver, which can be appllied to different domains (via generative techniques). It has in particular been used in the DiVA project to manage the dynamic variability of adaptive system by weaving aspects into models@runtime. It also inspired the GeKo weaver.