Zhiyuan Ouyang
I am a second-year Ph.D. student in Computational Mathematics at East China Normal University and Shanghai Innovation Institute, co-supervised by Assoc. Prof. Xiangyun Zhang and Prof. Junchi Yan.
My doctoral training is centered on generative models, computational mathematics, and AI systems. During my Ph.D., I was awarded the National Scholarship for PhD Students in China (Top 1.5%). I also received a provincial first prize in a collegiate C/C++ programming contest, which reflects my long-standing interest in combining mathematical rigor with practical system building.
My research focuses on theoretical understanding, generalization, and efficient controllable generation. In parallel, I work on large-scale AI systems including Slurm scheduling, DDP training, and practical deployment of visual generation pipelines. I am also interested in Web3 systems and prediction markets, especially on-chain execution infrastructure for complex strategy deployment around Polymarket.
I try to connect rigorous model understanding with practical engineering systems that are efficient, controllable, and deployable.
[Updated in 06/2026]

Working Experiences
Independent open-source execution router with multi-wallet routing, risk control, order aggregation, and MongoDB-backed state persistence.
Head of AI Department. Led controllable image generation pipelines, deployment planning, and team-level engineering standards.
News
- [06/2026] Open-sourced MimicPolymarket, an on-chain quantitative execution router for Polymarket.
- [05/2026] First-authored paper Primal-Spectral Generative Modeling accepted by ICML 2026.
- [03/2026] Awarded the National Scholarship for PhD Students in China.
- [2025] Published first-author work on thyroid nodule diagnosis in Computational Biology and Chemistry.
Publications
Zhiyuan Ouyang et al. (first author). ICML 2026 (CCF-A).
Zhiyuan Ouyang et al. (first author). Computational Biology and Chemistry.
Zhiyuan Ouyang et al. (first author). Communications in Statistics.
Technical Skills
- Programming: Python, TypeScript, Rust, C++
- Deep Learning: PyTorch, JAX, Diffusers, TensorBoard
- Distributed Training: Slurm, DDP, large-scale scheduling and performance tuning
- Web3 Infrastructure: Ethers.js, Viem, Polymarket CLOB SDK, Account Abstraction, MongoDB, Docker