PhD student, University of Hong Kong
1 paper at NeurIPS 2025
We propoed the first application of RLVR to directly enhance LLMs' proficiency in optimization modeling: SI-RL, achieves state-of-the-art on diverse public benchmarks