Dr Jerome Kieffer, originally physical chemist, led an in-silico laboratory in the pharmaceutical industry to predict chemical reactivity and analytical properties before joining the European Synchrotron light source (ESRF) in Grenoble (France). He is in charge of the online data analysis at the synchrotron where he develops software pipelines based on Python and running in real-time to analyse X-rays diffraction and absorption data. Python, thanks to its great flexibility and its ease of learning, is the prefered programming language for scientists who see it as a nice scripting language and a Matlab replacement. Because of the real-time constrains and the big data produced by synchrotron light sources, NumPy is often not fast enough to cope for such data deluge. A large amount of work is devoted on parallelization of algorithms in order to unleash the computing power of GPUs while keeping the easy-to-use pythonnic interface.