Description

Spécialisé dans la construction la modification et la maintenance d'écosystèmes data (data stack), sur le cloud ou bien sur des servers on premise, je peux vous accompagner dans la conception d'un écosystème intégré de données en fonction de vos besoins, en répondant à des critères de robustesse, d'efficience, d'intégration et légaux.

Langues

Français
Bilingue ou natif
Anglais
Bilingue ou natif
Portugais
Bilingue ou natif
Polonais
Bilingue ou natif

Préférences en matière de lieu de travail

En télétravail uniquement

Travaille majoritairement à distance

Warner Bros Discovery
AI Engineer
CINÉMA & AUDIOVISUEL
janvier 2025 - août 2025 (7 mois)
Paris, France
Custom MCP Server for Task Automation
Engineered and deployed a custom MCP server from scratch to empower non-technical users to execute complex data queries and automated tasks directly from their AI chat client.

Architected and built a modular, plugin-based backend using FastMCP (deployed with FastAPI on GCP), enabling to rapidly integrate new custom tools and capabilities with minimal overhead.
Developed the full-stack solution independently, from the server logic to the tool definitions, providing a seamless and powerful extension to the company's existing AI toolkit.
Automated internal workflows, such as database querying and information retrieval, to reduce manual effort and provide instant data access for team members.
Tech Stack: Python, FastAPI, Model Context Protocol (MCP), FastMCP, GCP

Modular Multi-Agent System with Human-in-the-Loop (HITL)
Designed and built a sophisticated, stateful orchestrator agent using LangGraph that dynamically routes user requests to a configurable suite of specialized sub-agents, handling everything from database interaction to external API calls.
Engineered a hierarchical architecture, personally developing the specialized sub-agents (e.g., SQL query, YouTube data retrieval) that were orchestrated by the primary agent.
Implemented a Human-in-the-Loop (HITL) workflow that pauses graph execution for user validation on ambiguous tasks (at the orchestrator level and subgraph level)
Integrated the complete agentic system with a Chainlit frontend to deliver an interactive chat application, managing the full communication pipeline from UI to backend.
Designed the system to be highly extensible, allowing new agentic capabilities (sub-graphs) to be added via simple configuration file changes, promoting rapid development and scalability.
Tech Stack: Python, LangGraph, LangChain, Chainlit
LangGraph Langchain Google cloud Python FastAPI
Warner Bros. Entertainment
Senior Analytics Engineer
DIVERTISSEMENTS & LOISIRS
avril 2020 - janvier 2025 (4 ans et 9 mois)
Paris, France
Building the international data ecosystem. From data sourcing to their reporting through dashboards.

Enabled up-to-date movie awareness tracking, powering Machine Learning models for dynamic Box Office predictions and campaign adjustments based on audience behavior.
Streamlined digital campaign monitoring (from Google and Meta Brand lift surveys), facilitating swift performance enhancements and budget optimization.
Leveraged Google search data and campaign outcomes to refine Box Office forecasting, aligning marketing strategies with real-time trends.

Technical:
Developed foundational code for Data Engineering, ensuring checkpoints, fallbacks, and retries in all ingestion processes.
Implemented data models using DBT to meet the diverse data needs
Created robust APIs, thoroughly tested, documented, and following design patterns, for seamless data ingestion and delivery.
Built an MLOps stack to ensure model reproducibility, facilitate model comparison, and streamline model deployment.
Conducted data landscape analysis to identify opportunities for data structuring and optimization.
Performed data transformation and integration from various sources such as emails, Excel files, APIs, and scraping.
Orchestrated data ingestion using Airflow, ensuring smooth and efficient data workflows.
Optimized queries and databases, particularly utilizing MySQL, to enhance performance and query efficiency.
Developed user-facing applications using Flask for data upload and seamless integration into databases.
Harmonized data and promoted awareness of available data resources throughout the organization.
Created dashboards using Apache Superset, tailored to specific use cases and integrated new functionalities as required modifying the source code.
Integrated ML models into dashboards, providing actionable insights and enhancing decision-making processes.
superset Airflow Apache Airflow SQL Python
Simundia
Senior Analytics Engineer
CONSEIL & AUDIT
août 2021 - août 2021 (1 mois)
Paris, France
Installation de Apache Superset sur AWS EC2
Partage des bonnes pratiques de maintenance et de modification du code source de l'outil Apache Superset
Façonnage de KPI et vues à suivre pour gérer l'activité
Construction de Dashboard autour de ces propositions
Formation de l'équipe à l'utilisation de Superset
PostgreSQL AWS Metabase superset