Challenge Entry · 2025

TaproQA

An agentic retrieval-augmented question-answering framework developed for the IEEE IES GenAI Challenge.

Overview

TaproQA is an agentic retrieval-augmented question answering framework built for the IEEE IES GenAI Challenge. It combines retrieval over a knowledge source with agentic orchestration so the system can plan, retrieve, and synthesise grounded answers.

Approach

  • Agentic control flow that decomposes questions and routes retrieval steps.
  • Retrieval-augmented generation to ground answers in source evidence and reduce hallucination.
  • Designed to the constraints and evaluation criteria of the IES GenAI Challenge.

Gallery

Presentation & videos