I am a computer science Phd student at University of Southern California, in, where I am very fortunately advised by Willie Neiswanger. I obtained my Master's and Bachelor's degree at ShanghaiTech University. My research interests include reinforcement learning, reasoning large language models and agents. Please feel free to contact me at: upup.ashton.wang at gmail dot com ****& ****Google Scholar, Github, X, LinkedIn

Working Experiences

• Summer 2026 - Research Intern, Qwen
  • Stable and efficient RL post-training for large language models.
  • Hosts: Guoyin Wang (Qwen Pilot), Hao Zhou (Qwen)
• Winter 2025 - Research Intern, BespokeLabs
  • Agentic RL post-training acceleration via reward/environment curation.
  • Hosts: Mahesh Sathiamoorthy, Alex Dimakis, Greg Durrett, Shreyas Pimpalgaonkar
• Summer 2025 - Student Researcher, BluelightAI
  • Pre-training cross-layer transcoders over reasoning models.
  • Hosts: John Carlsson, Gunnar Carlsson, Jakob Hansen

Selected Publications

Release

Reasoning LLM

• Chiyu Ma, Shuo Yang, Kexin Huang, Jinda Lu, Haoming Meng, Shangshang Wang, Bolin Ding, Soroush Vosoughi, Guoyin Wang, Jingren Zhou. FIPO: Eliciting Deep Reasoning with Future-KL Influenced Policy Optimization. [PDF][Web][Github][HF]
• Shangshang Wang, Julian Asilis, Ömer Faruk Akgül, Enes Burak Bilgin, Ollie Liu, and Willie Neiswanger. Tina: Tiny Reasoning Models via LoRA, ICLR 2026. [PDF][Web][Github][HF][X]
• Shangshang Wang, Julian Asilis, Ömer Faruk Akgül, Enes Burak Bilgin, Ollie Liu, Deqing Fu, and Willie Neiswanger. Resa: Transparent Reasoning Models via SAEs, 2025. [PDF][Web][Github][HF][X]

LLMs

• Ollie Liu, Sami Jaghouar, Johannes Hagemann, Shangshang Wang, Jason Wiemels, Jeff Kaufman and Willie Neiswanger. METAGENE-1: Metagenomic Foundation Model for Pandemic Monitoring, 2025. [PDF][Web][Github][HF][X]
• Shangshang Wang, Ziyu Shao, and John C.S. Lui. Next-Word Prediction: A Perspective of Energy-Aware Distributed Inference, IEEE TMC, 2023. [PDF]

RL

• Shangshang Wang, Simeng Bian, Xin Liu, and Ziyu Shao. Neural Constrained Combinatorial Bandits, IEEE TON, 2025 (previously IEEE INFOCOM, 2023). [PDF]
• Shangshang Wang, Ziyu Shao and Yang Yang. Constrained Dueling Bandits for Edge Intelligence, IEEE TNSE, 2024. [PDF]

Awards & Honors

• Nov 2025: Tinker research grant from Thinking Machines Lab
• Jan 2025: Nebius Research Credits Program Compute Grant