Career Profile

With more than two years of experience of deploying machine learning models into production, my experience in embedded software development makes me a particular good fit for machine learning applications that require highly optimized runtimes.

Experiences

Senior Machine Learning Engineer

September 2021 - Now
Sonos, Paris

Machine Learning Engineer

September 2020 - September 2021
Sonos, Paris

Development and optimisation of real-time on device speech recognition libraries.

Machine Learning Engineer Intern

February 2020 - July 2020
IBM, Montpellier

Development of machine learning solutions for lung cancer detection on CT scans. - Research and development of a medical data generation pipeline using GANs - Nodule detection and segmentation with DL models (Unet 3D, Retina Unet) - Building a cloud solution for inference integrated with ITK-SNAP (visualisation tool)

Team: Power sytems, Cognitive system Lab
Internship supervisor: Jean-Armand Broyelle

R&D Engineer Intern

April 2019 - August 2019
IBM, Montpellier

Design and development of a massaive data acquisition pipeline on Power9 servers.

  • Integration of an FPGA accelerator with CAPI
  • Integration of Nvidia GPU
  • Study and development of Proof Of Concepts integrating FPGA and GPU with optimized data transferts for clients
Team: Power sytems, Cognitive system Lab
Internship supervisor : Jean-Armand Broyelle and Bruno Mesnet

R&D Engineer Intern

June 2018 - November 2018
Bull (Atos), Echirolles

Development of hardware accelerators (FPGA) for Deep Learning algorithms.

  • Development of RTL kernels (VHDL - Vivado and SDAccel)
  • VHDL components verification with GHDL and GTKwave using UVVM
  • OpenCL programming
  • Code generator development to support multiple networks architecture acceleration (Python and TCL)
Team: Big Data & Cybersecurity department (Extreme Big Data team - Hardware acceleration)
Internship Supervisor: Alban Bourge

Projects

Here are some projects I made during my academic path.

Global Gene Expression Analysis: Determine Hormone Signaling Activation in Human Breast Cancer Samples - This project was part of the Machine Learning course @EPFL and was performed with PhD students @EPFL Brisken lab. The goal was to be able to cluster patients according to some gene activations specific to their receptivity to some hormones. Best semestre project Award.
Epileptic Seizure Prediction on iEEGs Signals Using Machine Learning - This project was a semester project supervised by Reza Ranjandish @EPFL Microelectronic System Laboratory. The aim of this project was to explore ML algorithms to detect epileptic seizure using iEEG recordings in order to integrate them in a brain implantable device developed at the laboratory.
Exploring French National Trafic Injuries Data - This project was part of the Applied Data Analysis course @EPFL. The goal was to explore 10 years of data from the national traffic injuries dataset in France provided by ONSIR and to extract meaningful insights.
Compression of CNN for Embedded devices - The goal of this project was to perform a study of state of the art technics to compress neural networks in order to make them more efficient on low constrained devices. I made this project @GrenobleINP TIMA laboratory on the supervision of Stephane Mancini.

Skills & Proficiency

Python

VHDL & HDL

C & CUDA

Computer architecture

Machine learning