Freelancer profile translated to English.

Description

Agentic AI Architect | LLM Platform | Evaluation & Observability | GenAI on AWS | 10+ years

AI Architect & Lead Data Scientist — 10 years of experience.

I design and deploy end-to-end agentic Generative AI architectures, from LLM platforms to complex application workflows.

What I build:

• LLM API Integration & Orchestration (LangGraph, Langchain, LiteLLM, Bedrock)

• Advanced RAG Systems (Docling ingestion, OpenSearch, contextualization, evaluation)

• AI Agents & Multi-Agent Architectures (LangGraph, CrewAI, OpenAI SDK, Skills, MCP...)

• MLOps & AI Infrastructure (Langfuse, Ragas, DeepEval, MLflow, Docker, AWS)

I cover the entire cycle:

• GenAI Platform Design (LLM Gateway, vector DB, observability)

• Agent Development with Complex Business Workflows

• Multi-Model Strategy (OpenAI, Claude (Bedrock), Mistral)

• Multidisciplinary Team Leadership & Business Collaboration

On my complementary projects, I use and integrate agentic automation tools

(n8n, MCP, Semantic Kernel) to connect AI systems with business workflows and

external data sources.

Sectors: insurance, energy, automotive, healthcare, finance.

Co-author of a scientific paper in medical imaging (Springer).

Certified AWS AI Practitioner (Early Adopter) & Cloud Practitioner.

Industry field of expertise

Languages

French
Native or bilingual

Workplace preferences

Can work on-site

Paris (up to 50km)

Malakoff Humanis
Generative AI Senior Data Scientist
BANKING AND INSURANCE
September 2024 - Today (1 year and 9 months)
Paris, France
- Architect of the Generative AI platform on AWS (Bedrock, SageMaker, LLM Gateway, OpenSearch, observability, evaluation).
- Development of a modular agentic architecture for business GenAI applications (LangGraph/Langchain orchestration, complex workflows, Tools, MCPs, Guardrails).
- Advanced RAG modules: structured ingestion (Docling), vectorization (OpenSearch), contextualization, pre-filtering, evaluation (Ragas, DeepEval, Langfuse).
- Advanced evaluation module as a microservice (agentic evaluation, RAG, tools)
- Leadership of a multidisciplinary team + collaboration with business teams.

Technical Stack: Python, FastAPI, SQL, MCP, Docker, Dataiku, Jenkins, AWS (SageMaker, Bedrock, Jumpstart..), LiteLLM, Opik comet, Langfuse, MLflow, Linux, OpenAI, Claude, HuggingFace, Langchain, LlamaIndex, Docling, Embedding, Ragas, Jenkins, SQL/NoSQL, React.
Python Langchain LLMs MCP LangGraph
Yélé consulting
Data Scientist
ENERGY AND UTILITIES
July 2022 - Today (3 years and 11 months)
Paris, France
Experience and Achievements
• Internal Projects
− Optimization of power grid management (As good as or better than human).
o Technical solution combining GNN and RL for better scalability.
o Development of a Django application for the product.
o Participation in pre-sales.
− Development of an internal Generative AI tool customized for internal assets.
o Design of the tool's architecture (Data, DB, Packages, OS Model, application).
o Integration of the RAG solution (Retrieval-Augmented Generation).
o Development of additional layers to adapt to user needs.
− Decarbonization Platform.
o Development of a web platform to help companies decarbonize.
− Data & AI Referent:
o Response to AI and energy calls for tenders.
o Support for data consultants in their missions.
o Proposal of AI topics internally.
External Projects
− Development of a financial management application for charging stations.
o Role of AMOA (Assistance to the Project Owner) on this project.
o Responsible for the MySQL architecture and DevOps part.
− Pro Bono at Latitudes:
o Responsible for facilitating workshops on technology awareness.

Technical Stack: Python (Pytorch, Geometry), API, SQL, R, Shiny, Django, Django-Ninja, Flask, Docker, Gitlab,
AWS, Dash, Streamlit, Linux, Nvidia, LLM, Llama, Langchain, Vectordb, Mistral, RAG
LLM Langchain RAG Cloud AWS
CAPGEMINI
Data Scientist Analyst at Digiposte
HUMAN RESOURCES
January 2021 - July 2022 (1 year and 6 months)
Paris, France
- Managing and exploiting data from various sources, including structured and unstructured data, to generate valuable insights for businesses.
- Responsible for creating and producing reports and indicators that provide clear and concise information to stakeholders for decision-making purposes.
- Designing and developing machine learning (ML) and deep learning (DL) models to address various business challenges and provide innovative solutions to complex problems.
- Leveraging data-driven insights to answer business questions and contribute to the development of internal knowledge, leading to improved decision making and better business outcomes. Tools: Python (Numpy, pandas, tensorflow, keras....), R, Shiny, Agile, Jenkins, GitLab, Jira, Docker
Python Computer Vision Deep Learning