Gilles HACHEME, PhD

Research Scientist

InFashAI

InFashAI's goal is to build inclusive fashion AI technologies. I am working with Ai4Innov to make it a reality.  Here is the link to our Github repository. 

African languages translation system

I am actively working on a project aiming to build a robust african languages machine translation systems. In fact most of african languages are considered low ressourced languages. I am working with the Masakhane community and Takwimu Lab to gather translation data for languages such as Fon (to French). And to train machine translation systems using state-of-art algorithms. 

A doc2vec search engine (CORD-19)

I led a team from UpAfro to build a search engine capable to extract relevant information from scientific articles about COVID-19. This work was done through the participation to the CORD-19 Kaggle competition. Here is the link to our contribution:

A doc2vec search engine 

SentEco

Information is an important raw material for economic agents. The SentEco project aimed to build an economic growth indicator by using hundreds of thousands of newspaper articles available on the internet. Articles were extracted using web scraping techniques and a dynamic sentiment analysis was done with them. 

Breaking point in the electric car market in China

Using data about combustion  and electric cars sales over several years, the objective was to determine the breaking point of the market: that is the moment the market for electric cars will be equivalent to the one of combustion cars. Other data sources such as chinese government subsidies for the purchase of electric vehicles were included in the analysis. This is a prediction exercise that I and the team I led were able to perform with an econometric model. This work was done for PSA China. 

Dtect- A fake news detection tool

Construction of a "fake news" detection tool that we nicknamed "Dtect". I was teamed up with Quentin Liance. We were accompanied by Pépite Paca-Ouest and La Belle de Mai Incubator.

I was appointed with Quentin Liance to represent the University of Aix-Marseille at the Resume Project 2018 entrepreneurship competition (Resume Project).  Find below the presentation  we did during the Resume Project.


Pitch deck.pdf

My Master Dissertation:

Why employers post wages in job ads: mixing machine learning and econometrics 

Master dissertation HACHEME Gilles.pdf