Machine Learning Engineer

• Oct 2021 - Present

Machine Learning Engineer

• Nov 2019 - Oct 2021

TotalEnergie - Atome: Users segmentation - Hierarchical clustering. Non Intrusive Load Monitoring - Electrical power load monitoring and disaggregation. Cooperl: Anomaly Detection on time-series. TotalEnergie - Atome: Users segmentation - Hierarchical clustering. Non Intrusive Load Monitoring - Electrical power load monitoring and disaggregation. Cooperl: Anomaly Detection on time-series.

Artificial Intelligence Research Engineer

• Apr 2019 - Sep 2019

Research on interpretability of decision tree forest based models. Design of a differentiable forest estimator and interpretability based on adversarial examples. DRT/LIST/DM2I/SID/LI3A Research on interpretability of decision tree forest based models. Design of a differentiable forest estimator and interpretability based on adversarial examples. DRT/LIST/DM2I/SID/LI3A

Artificial Intelligence Research Engineer

• Mar 2018 - Aug 2018

R&D Study and development of new methods for subjective video quality assessment based on machine learning. We tested many types of models like SVR, ANN, CNN and LSTM. We also used 6 different datasets to evaluate the robustness of our models and other state of the arts metrics. Our paper presented at the HVEI 2019: Augustin Lemesle, Alexis Marion, Ludovic Roux and Alexandre Gouaillard, "NARVAL: A no-reference video quality tool for real-time communications", Human Vision and… Show more R&D Study and development of new methods for subjective video quality assessment based on machine learning. We tested many types of models like SVR, ANN, CNN and LSTM. We also used 6 different datasets to evaluate the robustness of our models and other state of the arts metrics. Our paper presented at the HVEI 2019: Augustin Lemesle, Alexis Marion, Ludovic Roux and Alexandre Gouaillard, "NARVAL: A no-reference video quality tool for real-time communications", Human Vision and Electronic Imaging Conference, IS&T Electronic Imaging 2019 Show less R&D Study and development of new methods for subjective video quality assessment based on machine learning. We tested many types of models like SVR, ANN, CNN and LSTM. We also used 6 different datasets to evaluate the robustness of our models and other state of the arts metrics. Our paper presented at the HVEI 2019: Augustin Lemesle, Alexis Marion, Ludovic Roux and Alexandre Gouaillard, "NARVAL: A no-reference video quality tool for real-time communications", Human Vision and… Show more R&D Study and development of new methods for subjective video quality assessment based on machine learning. We tested many types of models like SVR, ANN, CNN and LSTM. We also used 6 different datasets to evaluate the robustness of our models and other state of the arts metrics. Our paper presented at the HVEI 2019: Augustin Lemesle, Alexis Marion, Ludovic Roux and Alexandre Gouaillard, "NARVAL: A no-reference video quality tool for real-time communications", Human Vision and Electronic Imaging Conference, IS&T Electronic Imaging 2019 Show less

3rd project @Paris Digital Lab - Kidz

• Jan 2018 - Feb 2018

7-week project with a startup specialized in the estimation of popularity of brands for childrens. We built new visualization and prediction tools based on socials networks datas and machine learning models. - Prediction Model (Python, Sklearn) - Backend (Python, Flask, MongoDB, MySQL) - Frontend (Angular, D3)

2nd project @Paris Digital Lab - Smart Inspector

• Nov 2017 - Dec 2017

7-week project with the french Ministry of Labour, and more specifically with the DGT (Direction Générale du Travail - Directory General of Labour). We developed of a new digital solution prototype for the labour inspectors to help them in their every day tasks. The application is a PWA which allow inspectors to manage their visits, get informations, take notes, share notes and generate reports. Development of : - a backend server (Node.JS, CouchDB) - PWA front-end (ReactJS+Redux… Show more 7-week project with the french Ministry of Labour, and more specifically with the DGT (Direction Générale du Travail - Directory General of Labour). We developed of a new digital solution prototype for the labour inspectors to help them in their every day tasks. The application is a PWA which allow inspectors to manage their visits, get informations, take notes, share notes and generate reports. Development of : - a backend server (Node.JS, CouchDB) - PWA front-end (ReactJS+Redux, CouchDB, SemanticUI)

1st project @Paris Digital Lab - WeCheck

• Sep 2017 - Oct 2017

7-week project for a very young start-up, WeCheck, which wanted to create a brand new way to exchange business cards on trade fairs, with a connected object. Development of : - a backend server (Node.JS v8, Sequelize) - two mobile apps: one for clients, one for managers (React Native) - an administration interface (React.JS) Creation of a connected object based on Arduino, using LoRa network to communicate through LoRaWAN to the backend server. 3D printed case.