Projects

PEPER: prediction of the prosumption of renewable energy by learning (Ecole Polythechnique, CentraleSupelec) 2018-2020 Artificial intelligence techniques and heterogeneous data science

  • Predict the production of renewable electrical energy from forecasts and assess the quality of these forecasts. • Study user behavior in order to predict their consumption.

  • Compensate in real time the difference between production and consumption thanks to storage means adapted on the one hand, but also thanks to the adaptation of the consumption in production when possible.

  • Proposal of artificial intelligence, deep learning and deep reinforcement solutions learning.

  • Implementation and validation of algorithms in TensorFlow python.

  • Keywords: Python, data science, artificial intelligence, control, reinforcement learning, DRL.

STC: Telecommunications and Cloud Services (Nokia, Orange Labs, IRT SystemX) 2018-2019 Artificial intelligence and optimization solutions for 5G C-RAN

  • Develop telecommunications networks, to respond in particular to increased need for agility and performance.

  • Make radio access networks more agile and more efficient.

  • Develop new concepts for monitoring radio access for future networks mobile, and develop new software monitoring tools.

  • Keywords: Python, C/C++, artificial intelligence, optimization, cloud, reinforcement learning, DRL, OpenAirInterface, OPL IBM CPLEX, 5G.

DVD2C: Virtualisation and Cloud Services for CDN (Orange, Viotech, Resonate MP4, Telecom SudParis, IRISA) 2015-2018 Virtualization and optimization techniques adapted to CDN networking

  • New B2B service allowing the operator to sell/rent this offer to content operators. • Assurance of effective cooperation and collaboration between Box Telecom, NFV Cloud CDNs. • The control of the Cloud using SDN.

Keywords: Python API OpenStack OpenDaylight, virtualization, SDN, NFV, docker, optimization.

UP‐TO‐US: User-centric Personalized IPTV UbiquitOus and SecUre Services (Orange, Telecom SudParis) 2014-2015 UPTOUS, took place within the framework of the European collaborative project Eureka Celtic.

The goal of UP-TO-US is to develop novel IPTV solutions allowing the customization of services according to the crossed networks.

Various projects:

2014 SE, is a realization of a web application using Java JEE and SQL server.

2013 AI, is an artificial intelligence project. It aims at adapting deep learning techniques for optimization using python.

2013 DB, is a telephone network administration and management project. It deals with the construction and administration of an ORACLE database for the performance study of telephone networks.

2012 DA, is a realization of a distributed real-time application using AADL, CORBA, and Cheddar real-time simulator.

2012 RT, is a real-time graphical simulation of a print manager using tubes and semaphores in C programming language.

2012 OR, is an implementation of a graph coloring application using graph theories in C programming language.