☰ Menu
× Professional Experience Education Programming skills Publications Social Media Version française
Avatar

Maxime Portaz

PhD in Machine Learning/Deep Learning

Paris, France

webcontact@portaz.fr

github


Avatar

Maxime Portaz

PhD in Deep Learning and Computer Vision

Paris, France

webcontact@portaz.fr

github


Professional Experience Education Programming Skills Publications Social Media Version Française


Skills

  • Machine Learning Algorithms
  • Deep Learning
  • Predictive Modeling
  • Big Data and Transfer Learning

Technical Skills

Python Programming

90%

Pytorch

90%

C/C++ Programming

70%

OpenCV

70%

Work Experience

Research Scientist - Trax Retail
May 2020 - Currently

Trax Retail - Paris
Innovation Team member. Building new machine learning tools for retail.

LANGUAGES : Python, NodeJS
FIELDS : Machine earning, Deep Learning, Touch Sensor, multimodal data
TOOLS : Pytorch, OpenCV, Tensorflow, AWS

Research Scientist at Qwant Research
Nov. 2018 - May 2020

Qwant Research - Paris
Building a Image Search Engine, with Image/Image, Text/Image and Text/Text similarities.

  • Deep Learning
  • Images, text and multimedia embeddings
  • Multimodal representations
  • GPUs and Graphcore IPUs programming
LANGUAGES : Python, C++, Javascript
FIELDS : Image Indexing, Deep Learning, Natural Language Processing, images/documents and queries embeddings.
TOOLS : Pytorch, Vespa, Kubernetes, NVidia DGX, Microsoft Azure, Graphcore IPUs

Graduate Research Assistant
Sept 2015 - Oct. 2018

LIG - Grenoble Informatics Laboratory Research and Development on the GUIMUTEIC project, on Multimedia Information Retrieval. I mostly worked with Deep Neural Network on image classification, image search, video analysis and gesture recognition. I used Similarity Learning with Siamese Networks and Convolutional Neural Network for scene and gesture recognition.
I worked on :

  • Scene Recognition, Gesture Recognition with Deep Neural Networks
  • Image Search with Neural Networks and learning to rank methods
  • Big-data and Transfer Learning on small data-base
  • Data-base creation and annotation

Which allowed me to develop or improve the following competences :

LANGUAGES: Python, C++
FIELDS: Deep Learning, Image descriptors, Image representation
TOOLS: Pytorch, OpenCV, Tensorflow

I also taught for 3 years in Grenoble Alpes University C, Python, Algorithms and Unix.

I had three Interns during my time in the Laboratory :

  • Manon Schmitt - 2018 - Image Fusion, Image Masks
  • Matthias Kohl - 2017 - Image Representation Learning, Triplet Loss, Automatic Region Detection
  • Chi Xu - 2016 - Image Descriptor Extraction, Bag of Visual Word, distance to objects evaluation

Teaching Assistant
Sept 2015 - June 2018

Université Grenoble Alpes

2018 : Python Programming and Algorithmic
2017 : Introduction to Unix and C programming
2016 : Introduction to Unix and C programming
Research Engineer
July 2015 - August 2015

Pervasive Interaction Team - Inria Grenoble.

LANGUAGES : C++
FIELDS : 3D models, scene generation, data fusion
TOOLS : OpenGL

Research Internship
Nov. 2014 - June 2015

Pervasive Interaction Team - Inria Grenoble. Interactive Expressive Figurines for narrative design.

LANGUAGES : C++, Javascript(nodeJS)
FIELDS : Computer Vision, data fusion, objects tracking, multimodal data
TOOLS : Optical Flow, IMU sensors, OpenCV, Kinect

Research Internship
2014

Prima Team - Inria Grenoble. Skeleton Based Imitation Game between human and humanoid robot (NAO)

LANGUAGES : Python, C++
FIELDS : Robotic, Computer Vision, 3D visualization
TOOLS : Blender Game Engine, NAOqi, Kinect SDK

Software Engineer Internship
2013

Prima Team - Inria Grenoble. Clustering and classification on multimodal data for intention detection in human-robot interaction.

LANGUAGES : Python
FIELDS : Machine Learning, supervised and unsupervised learning
TOOLS : Scikit-learn


Education

Université Grenoble Alpes
2015 - 2018

PhD Student - Multimedia Information Retrieval


Ecole Polytechnique - Paris
2017

Data Science Summer School


Computer Vision Summer School
2016

Sicily - ICVSS 2016


Ensimag - UGA
2013 - 2015

Master Degree - MOSIG Master Of Science In Grenoble - Graphics, Vision and Robotics


Université Joseph Fourier
2011 - 2013

Bachelor Degree - Mathématiques et Informatique



Programming Skills

Favorite Languages

Python

90%

C/C++

80%

Golang

60%

CUDA

60%

C#

50%


Favorite Tools

Pytorch

90%

OpenCV

80%

Unity

60%

Blender

50%

Caffe

40%

Operating System

Linux

90%

Windows

90%

Android

60%


Publications

2020

Graphcore C2 Card performance for image-based deep learning application: A Report

Ilyes Kacher, M. Portaz, Hicham Randrianarivo, Sylvain Peyronnet

QISS : an open source image similarity search engine.

Demonstration video
M. Portaz, Ilyes Kacher, Hicham Randrianarivo, Adrien Nivaggioli, Sylvain Peyronnet
2019

Image search using multilingual texts: a cross-modal learning approach between image and text.

M. Portaz, Hicham Randrianarivo, Adrien Nivaggioli, Estelle Maudet, Christophe Servan (LIUM), Sylvain Peyronnet (ELM)
2018

Information Access in mobile environment for museum visits - Deep Neural Networks for Instance and Gesture Recognition.

M. Portaz, PhD thesis

Object instance identification with fully convolutional networks,

M. Portaz, M. Kohl, G. Quénot, J-P. Chevallet, P. Mulhem, In Multimedia Tools and Applications (MTAP) 2018
2017

Fully Convolutional Network and Region Proposal for Instance Identification with egocentric vision,

M. Portaz, M. Kohl, G. Quénot, J-P. Chevallet, In IEEE International Conference on Computer Vision (ICCV) 2017

Figurines, a multimodal framework for tangible storytelling,

Demonstation video
M. Portaz, Maxime Portaz, Maxime Garcia, Adela Barbulescu, Antoine Begault, Laurence Boissieux, Marie-Paule Cani, Rémi Ronfard, Dominique Vaufreydaz, International Workshop on Child Computer Interaction (WOCCI) 2017

Construction et évaluation d’un corpus pour la recherche d’instances d’images muséales,

M. Portaz, J. Poignan, M. Budnik, P. Mulhem, J-P. Chevallet, In CORIA 2017 p. 17-34)

A system for creating virtual reality content from make-believe games,

A. Barbulescu, M. Garcia, A. Begault, M. P. Cani, M. Portaz, A. Viand, D. Vaufreydaz In Virtual Reality (VR), 2017 IEEE p. 207-208.

Making Movies from Make-Believe Games,

A. Barbulescu, A. Begault, L. Boissieux, M-P. Cani, M. Garcia, M. Portaz, V. Alexis, H. Pierre, D. Romain, R. Ronfard, D. Vaufreydaz. In WICED 2017.
2016

MRIM-LIG at ImageCLEF 2016 Scalable Concept Image Annotation,

M. Portaz, M. Budnik, P. Mulhem, J. Poignan,
In CLEF2016 Working Notes,2016.

Etude préliminaire à la recherche de photographies muséales en mobilité,

M. Portaz, P. Mulhem, J-P. Chevallet, In CORIA 2016, p. 335-344.
2015

The Grenoble system for the social touch challenge at ICMI 2015,

V-C. Ta, W. Johal, M. Portaz, E. Castelli, D. Vaufreydaz, In Proceedings of the 2015 ACM on International Conference on Multimodal Interaction. ACM, 2015. p. 391-398.