Alexis Gobé
Personal webpage

I am currently a R&D engineer in the SED (Service d'Expérimentation et de Développement) team at Inria, Sophia-Antipolis. I am currently working as tech leader of the nanophotonics platform DIOGENeS. Find below more information about myself as well as a sample of my past work. You can find links to contact me at the bottom of this page.

Since 2022

Research Engineer at SED at Inria, Sophia-Antipolis

Specialized in scientific computing & visualization, parallel computing

2020-2021

R&D engineer at Inria, Sophia-Antipolis

• Tech lead of DIOGENeS software suite

• Workshops coordinator and organizer

• DIOGENeS website maintainer

2016-2020

PhD in applied mathematics and numerical simulation at INRIA

Topic: Discontinuous Galerkin methods for the simulation of multiscale nanophotonic problems with application to light trapping in solar cells

2015-2016

Master degree in applied mathematics at Université de Reims Champagne-Ardennes

Topic: Modélisation Mathématique pour les Sciences de l'Ingénieur

PV simulation on hybrid meshes

I have used THYMuS in the context of solar cells and have shown the efficiency and benefits of using such configuration.

Multiscale features

I have worked on some realistic ultra thin film solar cells that involve multiscale features (nanometers layers surrounded in tens of microns ambient media).

Light-trapping simulation in realistic solar cells

I have developped a preprocessing tool to built unstructured fully tetrahedral meshes from in-situ AFM solar cell topography data. Those meshes were used to simulate light trapping and validation was done against laboratory measures.

Design optimisation

I have develop an optimization workflow using DiceOptim library, which is based on statistical learning methods (see EGO / Bayesian optimization), with DIOGENeS. I have done several optimizations, such that finding the best texturation in order to maximize the absorption capability of a device.

MHM-DGTD for Maxwell's Equations

I have developped a Multiscale Hybrid Mixed (MHM) solver for the 2D time-domain Maxwell's equations. I have also developped a coarse/fine two level mesh generator in prevision of a 3D MHM solver implementation.

Advanced mesh partitionning

I have proposed and developped advanced strategies to reduce load balancing for complex configurations (local refinement, hybrid meshes, local observables). An apriori estimation of the computation cost for each cell/face depending of the simulation parameters allows to change in accordance the nodes weight.

Hybrid mesh generator

I have developped from scratch THYMuS, a Toolbox for HYbrid MeSh, which helps building complex hybrid non-conformal meshes.

DIOGENeS software

From 2016 to 2019, I've been contributing in the DIOGENeS platform. Since 2020, I'm the tech lead of DIOGENeS.