Hi, I'm shy - huiyang shao (邵慧杨)

AI researcher and builder, obsessed with the fundamental principles behind everything — from optimization theory to pixels.

PhD in Operations Research at Tsinghua University, studying non-convex, non-smooth stochastic optimization. Previously earned my master's at the Institute of Computing Technology, CAS, where I received the CAS President's Scholarship (top 1%).

Author of 10 top-tier papers across NeurIPS, ICML, CVPR, ICCV, AAAI and TPAMI — including a NeurIPS Spotlight and an AAAI Oral — spanning AUC optimization, long-tailed learning, and diffusion acceleration.

Worked on post-training and distillation acceleration for Seedance at ByteDance Seed, knowledge graphs at Baidu, and user modeling for TikTok content safety. Now building AI-native products focused on person/group context wiki.