Robust Multi-Source Retrieval and Verification for Multi-Agent QA Systems

This project will strengthen an AI question-answering system by improving how it finds and checks information. It focuses on building a retrieval pipeline that can search across multiple document collections and then filter the results to keep only the most relevant pieces. She will also design a “verification agent” that double-checks whether each answer is really supported by the retrieved evidence and flags possible mistakes. The project adds a robust information layer to an existing multi-agent AI framework, deepening expertise in trustworthy and explainable AI. For the home institution, it offers concrete experience with end-to-end AI system engineering, from data pipelines to interfaces, and creates a basis for future joint projects, student exchanges, and shared publications.

Faculty Supervisor:

Yassine Yaakoubi

Student:

Partner:

École nationale des sciences de l'informatique

Discipline:

Engineering

Sector:

Education

University:

Concordia University

Program: