Benjamin Habié

Experience

• 

  LiveMentor

  Oct 2014 - Apr 2017

  Mentor

  Mentor dans le cadre de cours particuliers par internet en Mathématiques, Physique, Chimie et Anglais du niveau collège à BAC+2 et CPGE via la plateforme de cours en ligne LiveMentor.~150H

• 

  Eiffage Construction

  Jun 2016 - Jul 2016

  Stagiaire Assistant Conducteur Travaux

  Stage ouvrier de découverte de l'entreprise en première année, sur chantier. Travail en deux temps, une partie avec les ouvriers et assistant d'un chef d'équipe, l'autre en conduite de travaux.

• 

  John Paul

  May 2017 - Jul 2017

  Product Management Internship

  Apprentice engineer internship within the Product Management team. Monitoring of the production phase of websites and mobile applications, in dynamic collaboration with developers and sales department, from test version to final deployment.

• 

  Mobile Devices Ingenierie

  Apr 2018 - Aug 2018

  Cross Platform Mobile Development

  Cross platform mobile application development using Flutter SDK.App is meant to provide users with highly personalized and added value data as well as tailored supply of services for car navigation thanks to communication with OBD2 car technology through web agents and services.Tasks : - Research and development regarding the new technology Flutter- Creation of an extensive training course and a live-coding presentation conference regarding Flutter framework for other interns- GUI development in collaboration with the designer- MVC architecture pattern implementation- Cache implementation using Redux technlogyTechnologies : Flutter ; Dart ; React Native ; Docker Show less

• 

  Université de Toulon

  Sept 2018 - Jan 2019

  Deep Learning Research Project in Astrophysics

  R&D Lab internship working on a project in partnership with NASA (JPL) with data collected by WISE telescope (infrared survey telsecope).Bright comets detection through astronomical infrared multispectral images (FITS) analysis using Deep Learning algorithms.- Data annotation by JPL researchers- Preprocessing of raw data from WISE- Testing of several R-CNN (Region Convolutional Neural Networks) algorithms: Fast RCNN, Faster RCNN, Mask RCNN - Implementation of a multi-branch (one for each wavelength) Faster R-CNN model in Tensorflow (Keras)- Fine-tuning of the pre-trained fully-convolutional part on full (all wavelengths) FITS imagesTechnologies : Tensorflow ; Keras ; Astropy ; DS9 ; Faster R-CNN ; Siamese CNN Show less

• 

  Ecole Centrale de Marseille

  Sept 2018 - Jan 2019

  Music Generation with Generative Adversarial Networks

  Deep Learning application to automatic music generation - Project Leader- Extensive part on MIDI file preprocessing into understandable data Several implementations were studied and tested : - Dense & 1DConv VAE Autoencoder models trained on flattened Piano Roll data (2D representation of a MIDI file). The separated decoder part was used as our generator from the bottleneck dimension space (number of units in the last encoder layer)- Generative Aversarial Network (GAN) model implementation with LSTM layers trained on mini-sequences (typically one bar)Technologies : Tensorflow ; Keras ; Pretty-Midi ; Autoencoders ; GAN Show less

• 

  IBM

  Jan 2019 - Jan 2020

  Data Scientist @ IBM Watson

  * Extensive Computer Vision state of the art paper for internal use * Research project regarding automation of trauma (Hypoxic Ischemic Encephalopathy) detection on infants with Electroencephalogram (EEG) analysis :- Analyse EEG behaviors, limitations and characteristics of affected patients- Different signal processing tools : power spectral density on FFT of EEG, Morlet wavelets decomposition, Notch filters.- Clustering on labeled data for binary classification* Text classification for different mail buckets :- NLP multilayer perceptron model with TfIdf vectorization- Hyperparameter fine-tuning on NLP text classifier in production- Continuous Integration with regular additional feeding* R&D for text classification and text extraction on unstructured data :- Object detection pipeline for handwritten text with on images with modified Faster RCNN model for Scene Text Recognition- CRNN model with CTC loss for Handwritten Text RecognitionTechnologies : Tensorflow, Keras, IBM Cloud deep learning instances ; Microsoft Azure Show less

• 

  Usercube

  Jan 2021 - Jan 2024

  DevOps and Systems Engineer in IAM/IGA
• 

  Netwrix Corporation

  Jul 2022 - Jan 2024

  DevOps & Production Engineer
• 

  Qruise

  Jan 2024 - now

  Senior DevOps & MLOps Engineer for Quantum Computing

  Building entire cloud-native microservices infrastructures for quantum optimal control automation via machine learning, both on the cloud and on bare-metal.Working first-hand with quantum physicists and QPUs, with complex models and data that is highly labor-intensive and time-consuming to obtain, we require an extremely dynamic and modular architecture with everything controlled in-house and tailored to our very specific needs.