Consultant R&D Engineer

• Apr 2022 - Present

Crowdsourcing Project

• Jul 2019 - Jul 2021

Areas : R&D, Crowdsourcing, Data acquisition, Human-Machine Interface, Data science*Context :This computer research project aimed to develop some crowdsourcing algorithms and tools, an applicative experimentation was made on data from a citizen science (CS) program led by the Museum. The goal of the CS program is about to ask volunteers (the crowd) to participate on some species picture labeling tasks for biodiversity monitoring. *Goals :- Analyze the crowd behaviors during tasks annotations- Propose different picture assignment strategies to annote by user- Develop a tools for collecting and improving user experiences during the picture annotation tasks*Results :- Extraction and provision data about user profiling and picture features from the CS plateform- Transformation and integration of all data extracted to the HEADWORK plateform- Development of a dynamic and expressive interface for picture annotation- Experimentation and evaluation of the interface and assignment algorithm perfomance with real users*Technics :- Back-End & Front-End : PHP, html/css/Javascript - Database : SQL, MySql- Data exploration and statistics : Python (numpy, matplotlib, scikit-learn) - IA and Machine learning : Linear regression, Clustering (kmeans, HAC), NLP (spacy, nltk) - Crowdsourcing core : Automate structured model Show less

Web Application Project

• Jul 2019 - Dec 2019

Field : Web application, Computer science * Context : Creation of a new citizen science web platforms allowing volunteers to send their observations on nature. * Goal : - Ensure the recording of data sent by volunteers in the museum's database* Realization : - API development for data collection.* Techniques utilisées : - Back-End : NodeJS (Loopback 3), React-admin, Python- Base de données : SQL, MySql

Business Intelligence Project

• Oct 2018 - Jun 2019

Areas : R&D, Business Intelligence, Data Warehouse Spatial, Data Visualization, Cartography*Context :This computer research project aimed to build OLAP spatial systems on databases participatives, a study and application was realised on database from a citizen science program led by the Museum. The aim of the CS program is to mobilize farmers to share their agricultural practice for biodiversity monitoring.*Goals :- Propose new multidimensional analysis methods based on data with spatial information- Provide these analysis methods as a features on a interactive and user friendly proprietary geospatial tool *Results :- Establishment of an Spatial-OLAP system- Restitution of OLAP analysis results on the geospatial tool- Implementation of all retained visualization methods by the final users on the tool*Open source technics :- Programmation : Java - Database : SQL, Postgresql, PostGIS, Talend - OLAP : client (SAIKU), server (MONDRIAN) - Cartography : javascript - Template paradigm for the proprietary tool : JSON Show less

Artificial Intelligence Engineer

• Feb 2018 - Jul 2018

Areas : R&D, Computer Vision, Deep Learning, Software engineering * Context : Analysis and comparison about ancient book evolution over the different centuries of history * Goal : develop a new functionality on an image analysis tool for identify and caracterize differences between 2 images * Results: - optimization of image processing to caracterize all point of interest on image that considered as a "difference" - Build a deep learning algorithm to identify the differences types - Implementation of the comparison functionality on the image analysis tool * Technics : - Software interface : python-QT - Image processing : opencv - Image analysis : RANSAC algorithm - Feature processing : Textural analysis (GLCM), LBP - Machine Learning : CNN, SVM, Random Forest, Naïve Bayes, AdaBoost Show less

Data science researcher apprentice

• Jun 2017 - Aug 2017

Areas : Fundamental research, Recommendation system, Machine learning * Context : Improve user interaction and assistance during their analysis task on a proprietary tool * Goal : Recommend some request bundle to a user depend on his analysis interest * Results : - Extraction of "user interest" from all analysis queries recorded in the tool's log - Theoretical implementation of some "queries" recommendation algorithms based on similar and diverse analysis interest * Technics : - Data processing : python, R - Machine learning : KNN, K-means, BOBO (Bundle One By One) , CPS (Central-Plus-Satellite) Show less

Software developer

• May 2016 - Jun 2016

Areas : Software engineering, CRM, Web service * Context : Monitoring of computer equipment sales and stocks * Goal : Develop a system to interconnect an e-commerce plateform and a commercial managment tool in order to automate all inserting, deleting and updating process of product * Results : - Deployment of the e-commerce plateform and the commercial managment tool - Data integration from different supplier in a database - implementation of a service web (wsdl) allowing the connexion of the e-commerce plateform and the commercial managnment tool * Technics : - Web application : ASP.NET - E-commerce plateform : Prestashop - Server and database : OVH, MySql - Commercial managment : openSI Show less