Etienne Vincent

- Search-as-a-Service – Proof of Value: Designed an AI-powered search bar prototype for a funeral services company, benchmarked different engines (keyword-based, semantic, hybrid), and demonstrated value through improved result relevance and higher conversion rates. - Healthcare Data Architecture (HDS) – Engagement: Built an end-to-end ETL pipeline and Data Lake/Warehouse from scratch on GCP for a healthtech startup, with multi-source ingestion, GDPR/HDS compliance, analytical dashboards, and preparation for predictive use cases (e.g., insurance churn).

Designed an ETL pipeline to map network traffic flows (~100GB/day) and prevent ransomware attacks, including governance, data lineage, and a business rules engine. Led industrialization and R&D initiatives on graphs and Graph Neural Networks (GNN) for the detection of complex fraudulent networks.

En prestation à la Société Générale

Industrial Use Case – Predictive Maintenance of Wind Turbines: Developed anomaly detection models using IoT sensor data to predict degradation of critical components. Conducted ML modeling, validated key signals (frequency, spectral energy), and delivered results via notebooks along with client presentation and live demo.