Required skills
- Python
- Amazon Bedrock
- Amazon Bedrock AgentCore
- GitHub
- AWS
- Large Language Models
- Retrieval-Augmented Generation
- Agentic AI
- Vector Databases
Languages: French
Job description
Develop and optimize a Retrieval-Augmented Generation (RAG) system for processing and vectorizing brand plan documents using GenAI technologies.
Project overview
Context
At the client's, when a team is in the research and development phase for a molecule, other teams prepare the launch and commercialization plans for that molecule. They produce documents describing the launch strategy and plan (essentially PowerPoint brand plans), stored and referenced in product A.
A feature to leverage these documents has already been developed in A, but the current A agent also contains other data. The objective is to isolate this feature, simplify it, and generalize it so that other products can consume the brand plan data.
Missions
Technical Stack
Duration and Start Date
- Take over the existing system in A, in collaboration with the teams working on it
- Retrieve and ingest documents stored in A (essentially PowerPoints)
- Vectorize their content and convert it into the correct format
- Store this data in vector databases
- Build and optimize the RAG
- Create the agent, connect the client's reusable components, and expose it in the internal agent catalog
- Work in pairs with a full-stack developer responsible for data access segregation for the agent
- Python
- GitHub
- AWS, particularly Bedrock (LLM) and AgentCore
- General knowledge of LLMs, RAG, agentic approach
- Approximately 50 to 60 days of service (maximum 2 to 3 months)
- Start as soon as possible, ideally December 1st
Project details
- Start: flexible
- Remote: hybrid
- Location: Paris, France, FR
- Languages: French
Indexed by 9am's meta-search. You are a job provider? Claim or manage this listing
Found on collective.work