The 3rd year of the Magistère Ingénieur Economiste programme of study includes one graduation project called ESP, End-of-Studies Project. The purpose of an ESP is to enable students to carry out operational engineering data science work provided by our socio-economic partners and co-supervised by an academic from AMSE and the partner. A great way to check theory against real world applications, and a true transition to professional life. Djiby Balde, a student in Magistère Ingénieur Économiste and M2 Economics track Econometrics big data statistics, shares his experience.
Djiby, could you please tell us more about your work on ESP 2020?
For almost 5 months, I and 4 other students, had worked with the digital transformation department of Airbus Helicopters on a project whose overall objective was to build a document search engine. Our part focused on “text summarisation”, aiming to generate a summary of documents belonging to the same class (or same topic). Using several Machine Learning techniques, we first performed topic modelling to classify the documents by topic, before finishing with extractive and abstractive summarisation.
What lessons have you learned?
This was a pure data scientist project. I was very pleased to work on it and gain valuable experience. It was a long-term project, carried out over nearly six months in parallel with my studies. We were able to deepen our knowledge of Natural Language Processing/ Understanding (NLP/NLU) and to tackle new subjects. This allowed me to develop autonomy and to use knowledge learned in class. Working with others, I also learned how to take what each had to offer. In addition to the experience from my first internships completed in 2018 and 2019, the ESP gave me an understanding of how a company deals with problems using AI - Artificial Intelligencetechniques.
How did the defence go?
Despite the lockdown situation related to Covid-19, with my exams in progress, the first days of confinement had no real impact on me. However, all of us in the group were really concerned for ourselves, for families, team members and supervisors. Fortunately, AMSE and Airbus Helicopters set things up to ensure that the presentation went well and smoothly from home. We presented in such a way that people with no experience in data science could understand our work. I would like to take this opportunity to once again thank the entire AMSE team for investing their time and effort in this project, our academic supervisor Pierre MICHEL, and our professional supervisor Flavien RICHE (Data scientist at Airbus Helicopters – Amse alumnus), without forgetting the other students in the group.
© Photo by Aix-Marseille School of Economics