09/2014-07/2020, postdoc, supervised by me together with Frédéric Deschamps at the Institute of Earth Sciences, Academia Sinica, Taipei, Taiwan. Kensuke joined the lab of Robert (Bob) J. Geller at the University of Tokyo in 2006 as a fourth-year undergraduate, one year after me. We have always worked closely together, until 2020 when he decided to leave research. He defended his PhD in Tokyo in 2013 and then spent a year in Sang-Mook Lee’s lab at Seoul National University. Being extremely skilled in computing, Kensuke helped me a lot to refine my codes, especially my waveform inversion codes. All the numerical tools I had initially developed during my PhD were cleaned up during his. They are still used by students in Bob Geller’s lab and, later, in Kenji Kawai’s lab. Kensuke contacted me when he wanted to leave Seoul in spring 2014. I was then discussing a new joint project with Frédéric Deschamps (IES, Academia Sinica), who visited me in Paris, on reconciling geodynamics and seismology (which has interested me since the beginning of my career), so I asked him whether he would hire Kensuke and co-supervise him with me for a few years. His answer was a big “yes”. This was the beginning of our French-Taiwanese collaboration, which still continues. With Kensuke and Frédéric, we first applied my methodology of simultaneous waveform inversion for elasticity and anelasticity to image the base of the mantle beneath the Caroline plate. We published three papers on this subject, and he also took part in another project with Kei Hasegawa, a postdoc with Li Zhao at IES, Academia Sinica, which optimised spectral-element operators without increasing the computational cost. He came to IPGP twice for long stays: i) December 2014, to implement my attenuation inversion code in Java (funded by Academia Sinica); and ii) May-June 2018, to create the prototype of a new Direct Solution Method (DSM) code for a 3D planet using spherical harmonics (funded by IPGP). We were supposed to finish this project during his third visit in spring 2020 (I had obtained a one-month visiting researcher position), but the pandemic prevented us; the code is still under development, despite some contributions by Manuel Asnar over six months in 2020. Kensuke returned to Japan in August 2020 to start his new job as a software engineer.