Authors: A Nazari, S Sabouri, W B Zhu, R Jia, M Raghothaman, S Chattopadhyay Venue: IEEE Visual Languages and Human-Centric Computing (VL/HCC), 2026
Summary
A system bridging LLMs with CodeQL so novice programmers can interrogate large codebases. Uses RAG-based iterative query generation with self-testing. In a user study with 20 software engineers it increased task accuracy 3.7× while cutting completion time 32%.
Related
Thread: Human-AI Interaction in Knowledge Work — contributed as an educational research prototype. Thread: AI for Education — also part of the education thread.