Freelancer profile translated to English.

Description

Are your data an under-exploited asset?

Are your AI projects stuck between PoC and large-scale deployment?

As a PhD in Machine Learning and a freelance Senior Data Scientist & Machine Learning Engineer, I help organizations transform their data challenges into concrete, robust, and industrialized AI solutions.

With over 10 years of experience in major groups (Energy, Media, Consulting, Transport, Public Sector), I work across the entire AI project lifecycle.

✓ GENERATIVE AI (LLM, RAG): Development and deployment of solutions to automate, translate, enrich customer experience, and boost productivity.

✓ ADVANCED MODELING: Design of ML & DL models for prediction, complex time series analysis, graph mining, NLP (classification, anonymization), and Computer Vision.

✓ MLOPS & INDUSTRIALIZATION: Scalable production deployment on Kubernetes, Docker, and CI/CD pipelines (GCP, AWS, Azure DevOps).

✓ BIG DATA: Exploitation of massive volumes with Spark, Apache Beam, BigQuery, and high-performance cloud architectures.

CONCRETE AND MEASURABLE RESULTS:

✓ +96% recall on a recruitment model,

✓ 5x reduction in CVs to process,

✓ 81% accuracy for churn prediction,

✓ €500,000 in AI funding obtained for a major media group.

MY APPROACH: combining scientific expertise, product vision, and industrialization rigor to deliver AI solutions that move from the lab to the field.

As an independent consultant, I bring the flexibility, efficiency, and technical mastery needed to accelerate your digital transformation and achieve your strategic AI ambitions.

Languages

French
Native or bilingual

Workplace preferences

Can work on-site

Paris (up to 50km)

PWC France et Magreb
Sr. AI Engineer
CONSULTING AND AUDITS
March 2024 - December 2024 (9 months)
Neuilly-sur-Seine, France
Key Projects:
Project 1:Development of an AI for categorizing emails related to invoice reminders, thereby automating communication management and sorting.
Project 2:Creation of an AI to support recruitment, capable of accurately evaluating candidate profiles to simplify and accelerate the selection process.

Project Team:7 people (1 Data Director, 1 Team Lead, 2 Squad Leads, 4 Sr. AI Engineers)

Main Responsibilities:
• Analysis of business needs in collaboration with functional teams (Feature Teams).
• Scoping and planning of projects in conjunction with stakeholders.
• Development of AI solutions based on Large Language Models (LLMs).
• Organization and management of business data collection campaigns to train models.
• Evaluation of model performance: Recruitment AI: 96% recall for "high_match" profiles, 90% accuracy for "low_match" profiles.
• Design and development of high-performance APIs with FastAPI.
• Deployment of solutions on a Kubernetes cluster, ensuring scalability and robustness.
• Leading weekly meetings with business stakeholders to ensure project tracking and alignment.

Technical Environment:LangChain, Fast API, Python3, Pytest, MLflow, Sphinx, Kubernetes, Rancher, Azure DevOps, DSPy
Mathematical Models: GPT4/GPT4o
LeMonde
Tech Lead Data Scientist
November 2022 - March 2024 (1 year and 4 months)
Paris, France
Project:Development of Data Science models, mentoring, and team management
Project Team:7 people (1 PO/PM, 1 Data Director, 2 Data Scientists, 1 Data Engineer, 2 Developers)

Main Tasks:

- Gathering business needs from feature teams and participating in sprints (grooming, backlog).
- Development of a scalable crawler for data collection from archives of a live website.
- Implementation of pipelines: data quality monitoring, profiling, traceability, incremental calculation, periodic model training, and scoring.
- Design and supervision of a Data Science architecture on Google Cloud Platform (GCP) and cost monitoring.
- Development of predictive models:
- Early churner detection (Recall: 27%, Precision: 81%).
- Subscription propensity (Recall: 73%, Precision: 48%).
- Data analysis (A/B testing) and technical-functional documentation.
- Updating production infrastructure, writing unit/integration tests (Pytest, doctests), and post-mortems.
- Writing a proposal for GAFAM funding (€500,000 won).
- Integration of LLMs (GPT-3.5/4) for use cases: search engine (re-ranking), automated rewriting, translation (fine-tuning).

Technical Environment:
Technologies:Python, MLflow, Sphinx, GCP (Vertex AI, Kubeflow, Dataproc, Dataflow, BigQuery, Cloud Functions), Spark, Airflow, Scrapy, Jira, Azure OpenAI, Fastly, Cosmos DB.
Models: XGBoost, t-SNE, GPT-3.5/4.
Databases: BigQuery, CloudSQL (MySQL).
Enedis
Machine Learning Engineer
April 2022 - November 2022 (8 months)
Courbevoie, France
Project:Evolutionary maintenance and production deployment of a knowledge extraction application from verbatim feedback
Project Team:6 people (1 PO, 1 Scrum Master, 3 Machine Learning Engineers, 1 Web Developer)

Main Tasks:

Construction of a knowledge graph (PhraseNet) from verbatim feedback.
Restructuring the application into classes, packaging with Poetry, and code optimization (profiling and reduction of time-consuming functions).
Deployment of the application on a Kubernetes cluster with containerization via Docker.
Implementation of a REST API with Flask and documentation via Swagger.
Writing technical documentation and implementing DataOps practices to automate operational tasks.
Technical Environment:

Technologies:Python 3, Flask, Marshmallow, Sphinx, Kubernetes, Docker, JFrog Artifactory.
Model:Knowledge Graph.