About

Since 2018, I am a software engineer at Kalray, a semiconductor company specialized in the application of MPPA manycore technology to high-performance processing.

Between 2015 and 2018, I used to be in a postdoc position at the AriC project team, LIP, ENS de Lyon, France as an Inria engineer. My current research activities concern the conception of accurate, efficient, and reliable algorithms relying on the IEEE 754 floating-point arithmetic in a compilation context, notably ARM targets. In this scope, I look after the adequacy of integer and floating-point instructions as well as for the FMA exploitation. This work takes place within the Nano 2017 program headed by STMicroelectronics.

In 2014, I completed a PhD at the DALI research team of the Perpignan University on a source-to-source transformation dealing with (numerical) accuracy and (execution) time criteria. I developed a code synthesis aiming to automatically trade off between accuracy and time constraints. I thereupon worked for Numalis, a startup developing a spellchecker for calculations!

I am interested in computer arithmetics, computer architecture, compilers, performance measurements, and programming in OCaml.