Machine learning with a special focus on Deep Learning applied to time series (classification, clustering, prevision, generative models, signals desaggregation, predictive maintenance...).
Deeply involved in both methodological and technical sides. My on-going work deals with convolutional variational autoencoders (VAE) and generative adversarial networks (GAN), in the context of real-world use-cases at EDF.
Deep Learning technical environnement : Nvidia Titan X GPUs, Docker, Tensorflow, Keras, stack Python.
Data modeling with graph processing frameworks (Titan, Spark GraphX), in particular the whole french electrical distribution network.
GRP (Group Reporting Platform) is the AXA global mutualized reporting platform, handling high amount of data with high availability ; valuable strategical reports and ad-hoc analysis are provided across a wide variety of business and operating units.
Some key figures : 300 data sources (still growing), 30+ multidimensional cube analysis, 90+ millions of daily records, 20000 users over the world.
As a project manager, I was in charge of gathering business requirements from management teams and business unit leaders, cross-referencing user perception with my technical perspective.
Personally involved (design & specifications from scratch) in these strategic initiatives :
• Concentration risk management (advanced analysis and reporting)
• Group management accountancy (strategic reporting automation)
• Life insurance management (portfolios reporting and ad-hoc analysis)
Strong interest in technicals aspects, giving me the opportunity to hold a position of technical expert, within a context of high scalability, redundancy and availability in a strongly near real-time focused data warehousing platform.
Technical environment : Microsoft SQL Server (SSIS, SSRS, SSAS), Sharepoint with Excel Services, PowerPivot, custom software developments (C#)