Artificial intelligence for radiologists.

Deep convolutional networks for image recognition & analysis: classification, detection, segmentation.Deep Learning : Tensorflow, Caffe, TensorRT, Nvidia CUDA & DIGITSTechnologies : Docker, Kubernetes, GCloud, AWS, DjangoPlatforms : API, Linux on premises or embedded (Nvidia Jetson TX2)At deepomatic, we believe artificial intelligence should be made accessible to all. To achieve this goal, we have created a platform which enables businesses to develop their own image recognition systems. Vesta, an easy-to-use web platform, allows our customers to tackle their image-related problems by building and deploying custom AIs, without the need for a team of deep learning specialists. Customers simply upload their images, train their model through manual annotation and active learning and then deploy their API. We work with companies in a variety of industries such as automotive, healthcare, fashion and construction.

Providing help for a friend's Startup python / C++ / Fortran coding setup:- provide formations to command / hydrodynamics researchers on how to produce quality code (git, tests, CI...),- choose and install a computer for CI and simulation runs,- prepare project repositories for future open-sourcing.

Creation of a research app on mobile platforms for video quality perception tests.

Integration of MyScript's handwriting recognition SDK (in C) in a multi-threaded core application framework in C++ :- Reduce time to market for our customers wanting interactive and editable recognition- Anticipate, research and develop missing features to be added to the SDK (incremental recognition, heuristics for edition gestures, math solver, geometry).- Pool developments between desktop & mobile platforms (Win, Linux, MacOS, Android, iOS, WinRT, Tizen) for our applications (check out MyScript Calculator !).My skill set needed to accomplish this ambitious task :- Developer, mainly C++. Java (Android) & Objective-C- Multi-platform project management with Git and CMake- Unit testing, functional testing and performance testing with GTest- Continuous integration with Jenkins on MacOS, Windows, Linux, iOS plus java libs- JNI generation with SWIG for Java- OpenGL ES 2 rendering engine developed and integrated on all platforms- Writing Qt applications to directly test our application core on desktop (including OpenGL ES 2)I am currently the project leader on Geometry interactivity with a team of ~7, trying to reach the perfect handwriting IHM for several interactive shape use cases.Check www.myscript.com, MyScript Calculator (iOS, Android) and the www.webdemo.myscript.com

Participation au Startup Weekend sur le projet Menew Réalisation d'une partie du backend en Node JS, création de l'application Android,participation au front end en Jade et CSS. Le projet continue !

J'ai effectué un stage de 6 mois au Danemark à l'université d'Aalborg. J'y travaillait en tant que développeur C++ sur Uppaal, un logiciel de vérification formelle.