Summary
Xiaoze Liu is a PhD candidate and graduate research assistant at Purdue University with nine years of industry and research experience focused on post-training and alignment of large language models. His work—published at venues like ICLR and ACL and cited 1,000+ times—spans collaborative intelligence (mutual reinforcement learning, dynamic ensembling, federated preference optimization) and trustworthy compliance (reinforcement unlearning, agent-based defenses, knowledge-based calibration). He has interned at AWS AI Fundamental Research, Alibaba DAMO, and ByteDance, blending production-facing engineering with deep ML research. Based in West Lafayette, he combines a strong systems background from internships and a cross-cultural academic path (Zhejiang, Northeastern CN, Kanazawa) to uncover practical risks like model-merging supply-chain vulnerabilities while building mechanisms that make LLMs both more capable and more compliant.
9 years of coding experience
Bachelor of Engineering - BE, Computer Engineering, Bachelor of Engineering - BE, Computer Engineering at Northeastern University (CN)
Doctor of Philosophy - PhD, Electrical and Computer Engineering, Doctor of Philosophy - PhD, Electrical and Computer Engineering at Purdue University
Visiting Student, Computer Science, Visiting Student, Computer Science at Kanazawa University
Master of Engineering - MEng, Computer Software Engineering, Master of Engineering - MEng, Computer Software Engineering at Zhejiang University
Chinese, English, Japanese