Role and Responsibilities

• Develop and optimize reasoning and semantic search systems for AI agents including Kahina and Elyssa.
• Build and fine-tune LLM-based models to improve legal understanding and generate contextually appropriate responses.
• Design and implement vector database and Graph RAG architectures to enable efficient data retrieval and knowledge grounding.

Key Technologies and Tools

• Strong proficiency required in Python and NLP libraries/frameworks.
• Experience with LangChain, Azure AI Services, and AWS AI tools for model orchestration and deployment.
• Familiarity with vector databases and graph technologies such as Neo4J, FAISS, and Chroma.
• Expertise in Generative AI and prompt engineering practices for improved model outputs.

Expected Deliverables / Projects

• Implement a semantic search pipeline leveraging vector databases and graph structures for legal documents.
• Fine-tune and evaluate LLMs for legal-context understanding, including prompt engineering and response calibration.
• Integrate reasoning modules for agents (Kahina, Elyssa) to support contextual decision-making and retrieval-augmented generation workflows.

Requirements and Nice-to-haves

• Strong background in NLP, information retrieval, and applied machine learning.
• Experience designing RAG architectures, vector search indices, and graph-based knowledge representations.
• Experience with production AI services on Azure and AWS is highly valued.

How to apply

• Send your CV and a brief cover note to cv@e-tafakna.com including details of relevant projects, GitHub links, and availability for a 4–6 month internship.