Client case for BNP ITG
AI Smart Agent
Location
Paris
Project's functional area
Active participation in the study and implementation of a Smart Agent strategy. Support during the onboarding and adoption of AI tools.
Support for the development of the first AI models, providing AI expertise when required.
Support for the development of the first AI models, providing AI expertise when required.
Technological environment
Deep Learning / NLP
Docker / Kubernetes / MLflow / Python
Open-source LLM prompting (LLaMA, Mistral) on business use cases and domain-specific data
RAG: experience in implementing Retrieval-Augmented Generation using similarity search and Graph RAG
Deployment of LLM inference servers: TGI / vLLM
Docker / Kubernetes / MLflow / Python
Open-source LLM prompting (LLaMA, Mistral) on business use cases and domain-specific data
RAG: experience in implementing Retrieval-Augmented Generation using similarity search and Graph RAG
Deployment of LLM inference servers: TGI / vLLM
Methodology
SCRUM
Detailed description of the project
-
Active participation in B30 and in the implementation of a Smart Agent strategy
-
Development and maintenance of AI models
-
Business support for AI tool adoption
-
Code reviews
-
Knowledge sharing (technical and functional) and active participation in the AI Data Scientist community
-
Mastery of the test lifecycle
-
Documentation updates (technical and/or functional)
-
Contribution to the upskilling of local AI Centers of Excellence (CoE)