732 euro
6 years

My experience


ConsultingFebruary 2019 - Present

Mission 2: Celio CRM Marketing - Data Scientist

- SQL server
- Python
- R

Mission 1: Etia consulting - Deep Learning Engineer

Exploring Deep Learning applications as part of an internal need.
Image recognition on a Kaggle project "dog vs cat"). (Image classification according to two modes 'dog' or 'cat')

• State of the art on classical Deep Learning algorithms (Neural Network, CNN, RNN ...
• State of the art on Transfer Learning algorithms (Inception ResNetV2
VGG_16, Inception_V3, ResNet50, MobileNet, Xception)
• CNN classification: Keras

Technical Environment:

* Anaconda Python (Pandas, Keras, Sckilearn, Numpy, SciPy)

* Tensorflow

* Jupyter Notebook

IA SchoolFebruary 2020 - Present

AI teaching to engineering students (25 students) in IA School (Paris).

Practical course with Python projects:

- Data Analysis with Python (Pandas, Matplotlib, Seaborn, Beautiful soup for scrapping)
- Machine Learning: Regression & Classification Algorithm (Random Forest, KNN, Lasso, Ridge, Elastic Net)
- Deep Learning: LSTM, RNN, and CNN apply to Image and music Recognition
Transfert Learning with VGG16, GoogleNet, ImageNet, etc..

Simplon.coJuly 2019 - July 2019

Lead Data Trainer to Simplon.
Training session of 24 learners in the fields of data analysis and data science on Python, R and SQL (7-month course).

Data Analyst training:

* R language training

* SQL language training

* Python language training

* Data Analysis Training

Data AI program (Data Scientist)- Microsoft IA School:

* R language training

* SQL language training

* Python language training

* Data Analysis Training

* Machine Learning Training

* Deep learning Training

ADWAY - Cabinet de conseilApril 2018 - November 2018

BNP Paribas Risk ORC Solution (7 months):

The mission consists of the development of a risk control and risk assessment tool for Risk ORC Solution, an entity of the BNP Paribas Group. It follows a recommendation (warning) issued by the European Central Bank on the way in which the BNP's "Risk Control Self Assessment" (RCSA) were collected.

Creation of the app for the capture and evaluation of Risk s Control Self Assessment

Technical environment:
Excel, VBA and SQL

Lincoln FranceJune 2017 - April 2018

Mission 1: CSA Data Consulting - HAVAS Group (One month)

Marketing Project:
Measure of the media contribution on the company turnover under SAS.

SAS, VBA and Excel

Mission 2 : ORANGE - Intervention Unit

Data mining and Machine Learning projects:

• Forecast of the number of interventions for each of the territorial entities of Orange.

• Creation of a decision support tool to anticipate the payment of penalties to third-party operators (non-Orange) in the event of a default on an unbundled line.


• Prediction of the ”volume” variable: Statistical and econometric models (Exponential smoothing, HoltWinters, ARIMA) with HiveQL, Python (Pandas, Scikit-learn), R (forecast) and Excel

• Creation of a table of synthesis of the data under Hive including only the unbundled interventions (Wholesales) and prediction of ”Ranks” and ”Closing Code” variables using Random Forest Algorithm. (HiveQL and Python)

Business & Decision GroupFebruary 2016 - July 2016

I have made a final year internship of six months to Business & Decision. I was both researcher and data scientist.

On the research part, my mission was to measure the impact of the parametric variables on Machine Learning models.

On the data scientist part, from data provided by a rugby club, I had to find the probability that a player would be injured during a match or during a training.

I have also made a "State of the art" on Blockchain technology and it's applications (Bitcoins and more generally Crypto-currency) for the R&D department.

I resumed some of theses ideas in an article written for the Blog of the company. (French)

Obtained results:

• Have carried out a six-month research project including reading and synthesis of scientific papers and have written a 130-page research report.

• Have developed a preprocessing algorithm in R, resulting from this state of art to significantly improve the accuracy of a Machine learning algorithm. (Confidential)

• Have allowed to isolate the category of rugby player who are most likely to be injured in a match or at training (Clustering).

• Have carried out a research on the Blockchain technology and its applications on behalf of Business & Decision and have written an article which partially reproduce the results of this research for the blog of the company.

CNRS - National Scientific Research CentreJanuary 2015 - June 2015

The mean field theory of games was introduced in 2006 by Jean-Michel Lasry and Pierre-Louis Lions as the limit of non-cooperative games with many players. The main attraction of the theory of mean field game theory (Mean Field Games in English, MFG noted in the following) is the considerable simplification of interactions between players.

Players and determine their optimal strategy considering the evolution of the community (from the crowd of players) as a whole rather than all individual behavior (that is to say, each of the other players taken one by one ).

The MFG is well situated on the border between game theory (stochastic differential games to be exact) on the one hand, and the other optimization.

The main purpose of my intership was to write the two equations of a MFG game (Kolmogorov and Hamilton Jacobi Bellman) in the Poincaré Half plane.

LesaffreApril 2013 - July 2013

IT trainee.
Technical Achievements:
- Setting of an internal classification solution for technicals documents.
- Creation of VBA macros for engineers in the design office.
- Adapt previous VBA macros to make them compatible with the new version of Windows.

My stack

VBA, TensorFlow, Tableau Software, SQL Server, SQL, Scikit-Learn, SAS, RNN, Risk management, R Language, Python, Pandas, NumPy, NoSQL, Neural networks, MS Office, MS Excel, MongoDB, Microsoft SQL Server, Microsoft Excel, Matplotlib, Matlab, MapReduce, Machine Learning, LaTeX, Keras, Jupyter, Java, HTML, Hadoop, Deep learning, Data Visualization, Data analysis, D3.js, Clustering, C/C++, C#, Blockchain, Big Data, BeautifulSoup, Artificial Intelligence, Anaconda, Algorithms