Freelancer profile translated to English.

Description

Having started my curriculum with an MP preparatory class before joining Télécom Sudparis, I acquired a strong knowledge in mathematics which then led me to data science.

I started my first experience at IBM where I led numerous data science projects ranging from predictive maintenance, text mining, and image recognition. Then I joined the TNP teams as a senior Data Scientist to lead data science issues.

I carried out my first mission as a freelancer at Société Générale.

Languages

French
Native or bilingual
English
Fluent

Workplace preferences

Can work on-site

Paris (up to 10km), Paris (up to 10km)

Société Générale
Data Scientist Senior
BANKING AND INSURANCE
April 2020 - November 2022 (2 years and 7 months)
Paris, France
-Anomaly Detection: Identify abnormal behavior of a person on their expense reports.
Using behavioral analysis segmented by departments and expense report categories, we create statistical indicators (Z-score) to highlight individuals with atypical behavior.
We then present our results in a Dashboard format so that controllers can use our results more easily.

-Invoice Classification: Determine if an invoice has been assigned to the correct accounting account.
We determine the existing account classes and then predict the account class using invoice characteristics.
We then monitor the machine learning model using mlflow to track its performance over time, and the results are then presented in a Dashboard format.

-Prediction of intraday liquidity consumption of the bank
Behavioral analysis by counterparty, type of flow.
Real-time anomaly detection by family.
Docker Spark Airflow PostgreSQL
Tnp consultants
Senior Data scientist
CONSULTING AND AUDITS
September 2018 - March 2020 (1 year and 6 months)
Neuilly-sur-Seine, France
Technical supervision of the data science team.
Gathering business needs as well as functional and technical specifications.
Creation of AWS SageMaker datalab for development environment
-Development of a complete car dynamic pricing solution on the cloud integrating a web interface and an evolving machine learning model
Python Docker Amazon Web Services
IBM
Data scientist
CONSULTING AND AUDITS
January 2016 - September 2018 (2 years and 8 months)
Bois-Colombes, France
Predictive maintenance: Renault (6 months)
- Study of a car manufacturing line to improve machine efficiency.
- Sequential analysis of error, warning, and information messages to determine a recurring sequence preceding a breakdown. Grouping breakdowns into different labels, analysis with the business of the relevant lag for identification of relevant sequences.
- Production deployment of the selected sequences.

Energy overconsumption prediction: EDF (4 months)
- Enrichment of the database with meteorological data streams (IBM owning Weather Company)
- Development and deployment of the model to predict energy overconsumption.

Internal challenge: IBM (2 months)
- Use of graph theory on the RATP network to define the closest meeting point for 2 people.
-
Predictive maintenance: Faurecia (1 year)
- Study of car part manufacturing using parametric data.
- Detailed descriptive analysis of the performance of each factory, machine, and daily PDF reports sent by email.
- Detection of parametric anomalies on machines to alert operators.
- Optimization of parameters (temperature, pressure, etc.) using decision trees to improve part yield.
- Multi-parameter clustering to detect patterns preceding a breakdown.

Document classification: Ministry of the Armed Forces (6 months)
- Integration of several market solutions (Proxem, ESRI, etc.).
- Optimization of a first-generation search engine with NLP techniques for better reactivity/planning of officers.
Python Docker