Quang Liem Tran

Conducted research at Stanford’s STAI Lab on neuro-symbolic reasoning for multi-hop question answering, developing a context-mutated framework that adapts symbolic rule chains dynamically through contextual feedback. Designed a neural retriever, symbolic rule evolution engine, and verifier module, enabling interpretable reasoning and real-time adaptation under zero-shot settings. Achieved a 15% accuracy gain on HotpotQA compared to state-of-the-art neuro-symbolic baselines.