Freelancer profile translated to English.

Description

Data Scientist with 5 years of experience, I have led numerous personal and client projects in parallel with my missions and I am available for full-time or ad-hoc missions.

After 10 years spent in the simulation and calculation departments of automotive manufacturers and equipment suppliers, I retrained in data science 6 years ago.

After an initial permanent contract experience at Neossys, where I developed scrapers and designed the startup's first machine learning and NLP solutions, I chose to return to freelancing. I then assisted C-Ways on customer knowledge topics, developed anomaly detection algorithms for the RATP, and then worked at ArcelorMittal on forecasting and pricing projects.

More recently, I collaborated with Mediaperformance on marketing performance measurement, modeling, A/B testing, and the construction of decision-making indicators, as well as with Veesion, where I contributed to the evaluation and monitoring of machine learning models in production in a computer vision context.

In parallel, I regularly intervene at the Artefact data science bootcamp, where I lead two weeks of training, one dedicated to unsupervised machine learning and the other to NLP.

Finally, I have led several personal and entrepreneurial projects: an application for an advertising agency to build Google Ads campaigns through scraping and AI, a project to support e-merchants focused on performance and customer knowledge, and a SaaS for enriching B2C CRM data and generating personas using data science and AI.

I would be delighted to continue my career within your company and to leverage my data science expertise to address concrete and high-impact issues.

Industry field of expertise

Languages

English
Fluent
Italian
Fluent
French
Native or bilingual

Workplace preferences

Can work on-site

Metz (up to 50km), Paris (up to km), Luxembourg (up to 20km)

Enree.ch
Data Science / AI Application Development
E-COMMERCE
September 2024 - Today (1 year and 9 months)
Enreech is a SaaS solution that allows e-commerce companies and their partner agencies to better understand their customer base.
To achieve this, Enreech enriches and segments customer databases using external sources and segmentation algorithms.

Technical Stack and Architecture
Streamlit for the application's user interface.
Next.js for the homepage, landing pages, and blog
APIs developed in Python (Flask) and deployed on Cloud Run
Google BigQuery for storing, querying, and aggregating enriched customer data.
Pub/Sub for asynchronous messaging management.
Deployed on Google Cloud Run.
Custom customer segmentation algorithms developed specifically for this purpose.
Authentication and user management via Firebase Auth.
Generation of reports and persona descriptions via Gemini.
Key Features
Automatic enrichment of customer profiles by cross-referencing external and internal sources.
Advanced segmentation based on enriched data and creation of well-defined and rich personas.
Results
Developed in 3 months (evenings and weekends).
Several e-commerce brands already satisfied.
Google cloud Python Next.js Streamlit Gemini artificial intelligence Generative AI Python Flask
ArcelorMittal
Data Scientist
RAW MATERIALS INDUSTRY
April 2024 - Today (2 years and 2 months)
Luxembourg, Luxembourg
Within ArcelorMittal's CMO department, the Data Science team is responsible for developing models that enable sales teams to optimize customer pricing. It is also responsible for creating intelligent dashboards to facilitate business team decision-making. In this context, I am responsible for the dashboard dedicated to forecasting future prices, based on price indices. This dashboard integrates forecasts of these indices over a three-month period, offering strategic support to sales teams.
I also developed a forecasting package that allows for easy comparison and testing of different models, optimization of their parameters, and integration of exogenous variables from various sources. This package is currently in production and used daily by end-users.
Machine learning Time Series Forecasting Tableau software
RATP
Data Scientist
TRANSPORTATION
March 2023 - April 2024 (1 year and 2 months)
Paris, France
As part of the development of the internal application used by technicians to monitor train maintenance, RATP engaged data scientists to implement anomaly detection algorithms to identify interventions required among the numerous fault messages sent by trains.
After an exploratory phase analyzing the available data, I built an anomaly detection algorithm to identify the most relevant fault codes.
As part of improving and standardizing practices within the data science team, I participated in the development of an internal package enabling data scientists to analyze data and build models more quickly, robustly, and reproducibly. This allows for faster deployment of these models on new lines and new equipment.
Data science Anomaly Detection