Description

My professional journey began with a belief in the digital economy as the driving force of the modern world, with data as its pivotal fuel. My expertise centers on harnessing data and machine learning models to create value across diverse industries.

As a data scientist engineer, I'm adept at designing and deploying high-performance machine learning models. Much of my career has been dedicated to exploring how these models transform businesses, predicting market trends, optimizing operations, and enhancing customer experiences.

What sets me apart is my versatility across domains. I've collaborated with healthcare firms to improve diagnostic precision, assisted financial institutions in fraud detection and investment optimization, and streamlined manufacturing supply chains.

My approach hinges on understanding sector-specific needs and tailoring machine learning models accordingly. I thrive on tackling complex challenges and crafting innovative solutions that yield a sustainable competitive edge.

I specialize in constructing diverse data models—linear, non-linear, generalized, or mixed—employing modeling, estimation, testing, and diagnostic techniques. Identifying pivotal variables and testing hypotheses using advanced modeling techniques is my forte.

I interpret data to make predictions, considering their inherent randomness. Proficient in analyzing complex data, be it time series, text, networks, or images, I hybridize connectivist AI (neural networks) with symbolic approaches (reasoning engines), utilizing complementary technologies such as semantic web and multi-agent modeling.

In terms of technical skills, I excel in applied mathematics for data analysis and AI. Proficient in C, Python, and R, I work with relational databases via SQL and handle data lakes, particularly on Hadoop & Spark architectures.

My work encompasses data collection, preprocessing, and analysis. I engineer relevant variables for model training;

Langues

Français
Bilingue ou natif
Anglais
Bilingue ou natif

Préférences en matière de lieu de travail

Accepte de travailler sur site

Paris (jusqu’à 50 km)

Ville de Paris
Data Scientist
décembre 2022 - Aujourd'hui (3 ans et 6 mois)
Île-de-France, France
En tant que Data Scientist, j’ai conçu et déployé des modèles de machine learning et d’intelligence artificielle appliqués à des problématiques métiers variées : scoring de crédit, segmentation client, classification d’images, prédiction énergétique et analyse de données massives.

Mes missions incluaient :

le traitement et la valorisation de données structurées et non structurées,
la conception de pipelines de données avec Python, SQL, PySpark et AWS,
le développement de modèles de Deep Learning (VGG16, MobileNetV2, CNN, BERT),
l’optimisation et l’évaluation de modèles prédictifs,
le déploiement d’API et dashboards interactifs (Flask, Streamlit),
la mise en place de pratiques MLOps et de suivi de dérive des données.

J’ai également travaillé sur des architectures Big Data et cloud (Hadoop, Spark, AWS EMR) afin de gérer des volumes importants de données et automatiser les traitements distribués.

Cette expérience m’a permis de renforcer mes compétences en analyse statistique, intelligence artificielle, visualisation de données et gestion de projets data orientés business.
PySpark Python AWS SQL DevOps
Ville de Paris
Data Scientist
septembre 2023 - octobre 2023 (1 mois)
Île-de-France, France
Project Challenge : The start-up "Fruits!" aims to develop innovative solutions for fruit harvesting while preserving fruit biodiversity. In this mission, the company seeks to establish a data processing pipeline to create a first version of the fruit image classification engine.
ville
Data Scientist
avril 2023 - mai 2023 (1 mois)
7bis Rue de Lesseps, 75020 Paris, France
Missions: Assist Olist's teams in understanding different types of users by using clustering methods to segment customers on an e-commerce website, thereby enabling the Marketing team to target their communications more effectively. Work Done: Data Analysis:
• Analyze the distribution of the number of orders per customer.
• Identify and target the most interesting customers using the RFM (Recency, Frequency, Monetary) method.
• Enhance feature engineering for better data comprehension. Modeling and Clustering:
• Determine customer profiles for each cluster and ensure their relevance from a business perspective.
• Train clustering models, including K-means or DBSCAN, and fine-tune hyperparameters for accurate segmentation.
• Visualize the learning curve to assess model performance.
Temporal Analysis:
• Simulate multiple periods (T1, T2, T3) and visualize the evolution of the Adjusted Rand Index (ARI) to measure cluster quality over time.
• Measure cluster divergence using ARI to determine when the clustering model becomes obsolete and requires an update. Cluster Stability:
• Compare cluster divergence by cluster number to ensure the stability of clusters, especially those of valuable customers. Data Leakage Prevention:
• Implement measures to prevent data leakage, including dividing data into training, validation, and testing sets.
• Use cross-validation to evaluate model performance and ensure its robustness and generalizability. Maintenance Recommendation:
• Provide a recommendation for how often the segmentation should be updated to remain relevant. Coding Standards:
• Ensure that the code complies with the PEP8 convention for usability by Olist. Results Communication:
• Prepare a clear and actionable description of customer segmentation for optimal use by the Marketing team.