With a strong expertise in artificial intelligence and data analytics, I have had the opportunity to work on and lead end-to-end projects in various fields such as finance and healthcare, delivering value to each profession through specific use cases that seamlessly integrate into both business and technical workflows.
Lead and Develop AI and Data Analytics Initiatives for Natixis' Asset & Wealth Management Division. Key Projects:
Generative AI : Design and implement generative AI solutions within a highly secure environment, using cutting-edge technologies such as quantized Large Language Models (LLMs) and Retrieval-Augmented Generation (RAG) systems. The goal of the use case is to provide the business team with a GenAI tool that allows them to interact with internal databases using natural language.
Fund Performance Analytics Platform: Design a data analytics platform to monitor fund performance by integrating external and internal data. This includes a dashboard for in-house versus competitor fund comparisons and an email alerting system for key performance indicators.
Client 360: Develop data analytics projects that involve extracting, loading, and analyzing customer engagement data to construct comprehensive 360° dashboards. Our goal is to leverage data insights to support strategic decision-making and provide a 360 view of client engagement channels.
Co-founder and head of the AI department at the startup DataPathology, specialized in digitizing the processes of pathological anatomy and using AI approaches for cancer detection. I was responsible for leading and developing AI solutions, particularly for breast cancer.
My role includes the following:
Implementation of the scientific approach to address the cancer diagnosis problem
Implementation of the technical approach
Technical and clinical evaluation of the approach
Development and deployment of a web application to make the application available to doctors
Build Machine Leaning solutions for the CIB branch of Natixis. Key Projects:
Interest Rate Hedging: Building a tool that automatically helps manage the risks of interest rate changes for derivatives making smarter decisions by using recurrent neural networks to optimize the hedging.
Use NLP to analyze the minutes from client minutes, and turning their words into actionable insights
Build a cognitive platform for Information Retrieval tasks that improves our understanding of unstructured and semi-stuctured document, primarily using NLP & Computer Vision techniques. Key features of the platform include:
Named Entity Recognition to accurately extract specific entities from PDF documents
Entity Relationship Extraction
Deep Learning system for efficient table extraction and comprehensive document layout analysis.
Create a web API for the beta testing phase of the tool, facilitating user interaction and feedback.
IMT Atlantique, as a public top French higher engineering school combining digital technologies, energy, and environmental sectors, proposes engineering degrees mainly through programs at Master level.
Main courses:
Advanced statistics (Estimation, Classification and Decision )