About Haboubacar
Data Scientist – NLP and Gen AI Specialist
French
Native or bilingual
English
Fluent
Experience
- Caisse des Dépôts et ConsignationsData Scientist - EngineerPUBLIC SECTORMarch 2024 - Today (2 years and 5 months)Paris, France- Development of an AI algorithm to identify certifications related to the ecological transition, with a classification accuracy of 80% (CamemBERT Embedding + Cosine Similarity & CamemBERT Fine-tuning)- Semantic clustering of 25,000 certifications for document indexing (Cosine Similarity + Find-Union algorithm)- Development of a semantic file search engine on the data service's OneDrive (Langchain, Faiss, Camembert, Streamlit)- Design and deployment of a text document vectorization API (hugging face, Camembert, onnx runtime, FastAPI)- Design and optimization of data pipelines via Alteryx and PySpark: automatic data provision in response to business requests- OpenData technical lead: creation, publication, and maintenance of data set feeding pipelines- Conducting project presentation workshops with business teams
- Île-De-France-MobilitésData ScientistTRANSPORTATIONOctober 2022 - March 2024 (1 year and 5 months)Paris, France- Design and deployment of a RAG on the functional & technical specifications of the IVTR team- Design and deployment of an AI Agent (interface with GPT-4) to simulate transport data for testing- Creation of a Flask webApp for data flow generation (chat interface)• Deployment of the webApp on Azure (App Services, Azure Functions) and implementation of a CI/CD- Automation of non-regression tests, reducing manual execution time from 5 days to 3 hours, a 90% productivity gain- Conducting non-regression tests for the IDFM route search application, optimized for the Paris 2024 Olympics- Writing specifications, managing anomalies, and technical monitoring of the IVTR handover- Facilitating agile instances and coordinating connection tests for new partners
- PwC FranceData Scientist NLPCONSULTING AND AUDITSOctober 2021 - August 2022 (10 months)Paris, France- Training and production deployment of eight email classification models on Databricks via MLflow (comparing CamemBERT, SVM, and Perceptron models)- Design and implementation of a robust MLOps pipeline, integrating a feedback loop for continuous evaluation and data drift detection- Creation of an UiPath robot interfacing the email classification model and the feedback loop, automating email classification in the customer inbox- Development of a high-performance archiving recommendation system to identify and eliminate obsolete documents, resulting in a 20% reduction in storage space used- Collaboration with development teams and project managers to align solutions with client needs.
Recommendations
Be the first to recommend Haboubacar
Help this freelancer shine by sharing your experience working together.
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
- Engineering Degree, Artificial IntelligenceCY Tech2022Statistiques - Machine learning - Deep learning - Deep learning - NLP
- CPGE MPSI-MPCPGE Lycée Moulay Idriss (Fès-Morocco)2019CPGE option Maths-Physique