Freelancer profile translated to English.
Back to original languageAbout Stéphane
Most AI projects stop at the same point: a prototype that works locally, and no one to turn it into a system the company can actually use. That's precisely where I come in.
Senior Data Scientist with 6 years of experience designing and deploying ML and GenAI systems. I cover the entire chain - modeling, development, deployment, security, observability - without handoffs between stages.
1) MedRAG - RAG system deployed on GCP Cloud Run: Private IAM-only FastAPI API, Streamlit UI, BGE embeddings, Chroma index, page-by-page cited sources, three compartmentalized service accounts, 24 pytest tests.
2) FoodHub - Industrialized AI backend on AWS: FastAPI, deterministic repository layer preventing direct LLM SQL access, private RDS PostgreSQL, ECS Fargate, MCP server exposing controlled tools, 4-stage CI/CD pipeline.
Public code, architecture, and runbooks on GitLab.
Hybrid work suits me, but full remote is my preferred mode.
French
Native or bilingual
English
Fluent
Can work on-site
Strasbourg (up to 50km), Paris (up to 75km), Mulhouse (up to 10km), Metz (up to 10km), Lyon (up to 20km)
Experience
- entrepreneurial GenAIFounderNovember 2025 - Today (9 months)Strasbourg, France• • Scoping a B2B offering focused on reliability and auditability of documents in sensitive inter-company contractual exchanges.• • Technical prototyping of a RAG system exploring retrieval architectures, the methodological framework for document reading, and traceability constraints.• • Conducting field interviews with delivery managers, COOs, and contract managers to qualify the problem and identify operational use cases.
- ENYGMA SASData Scientist & ML EngineerJanuary 2023 - January 2025 (2 years)Strasbourg, France• • Design and production deployment of fraud detection and transaction scoring models (success/failure/gray area decisions) on GCP, scoring real customer transactions.• • Batch and real-time pipelines (MySQL, SQLAlchemy, Ray) with scoring latency between 500-600 ms.• • Comprehensive MLOps practices: Docker, GitLab CI/CD, deployment on Vertex AI / GCP, automated model KPI monitoring.• • Customer segmentation (RFM, behavioral clustering) and anomaly detection for proactive fraud prevention.• • Explainability (SHAP, LIME) for transparency, auditability, and regulatory compliance; writing the company's five-year CIR report.
- FM LOGISTICData Scientist & Data Project ManagerJanuary 2020 - January 2023 (3 years)57370 Phalsbourg, France• • Churn and productivity predictive models: leading to a 27% reduction in operational costs.• • Forecasting models (time series) for country budget projections, achieving 10% higher accuracy than manual reference methods.• • Containerization and deployment of ML services (Flask, Docker) in a microservices architecture; automation of business processes (Python, UiPath) generating an estimated saving of 2 FTEs.
Recommendations
These freelancer profiles also match your criteria
Agatha Frydrych
Backend Java Software Engineer
4.7
(3)
2
Baptiste Duhen
Fullstack developer
4.6
(4)
5
Amed Hamou
Senior Lead Developer
4
(2)
7
Audrey Champion
Web developer
4.3
(3)
4
Education
- Post Graduate Program In Artificial Intelligence and Machine Learning : Business RecommendationsGreat Learning & TEXAS McCombs2022Post Graduate Program In Artificial Intelligence and Machine Learning : Business Recommandations
- Business School Diploma – Financial ControllerEM STRASBOURG BUSINESS SCHOOL2018Business School Diploma – Financial Controller
Certifications
- Artificial Intelligence and Machine Learning ExpertUniversity of Texas at Austin2022