AI Researcher / Software Engineer / Builder
Building reliable AI agents and ML systems.
I'm pursuing a B.S. in Computer Science at the University of Southern California ✌️
My research at USC and in open source focuses on post-training, data-centric ML, and self-improving agents. I am fortunate to work with Yue Zhao and Jieyu Zhao at USC, and this summer I am joining Kimi (Moonshot AI) as a technical staff intern.
Reach out to me at siuw [at] usc [dot] edu
Selected Projects
View project
Monolith
Infrastructure for recursive language models, enabling AI coding agents to reason persistently across arbitrarily large contexts without context-window rot or lossy summarization.
Multi-agent framework for automating anomaly detection across different data modalities using LLMs.
Open-source initiative curating datasets for training agentic LLMs.
Everyday menu of USC dining hall - an mobile app helping USC students easily access daily dining hall menus and food locations.
Published Papers
AD-AGENT: A Multi-agent Framework for End-to-end Anomaly Detection
2025
T Yang*, J Liu*, W Siu*, and 6 more
Findings of IJCNLP-AACL, 2025
Pyod 2: A python library for outlier detection with llm-powered model selection
2025
S Chen*, Z Qian*, W Siu*, and 8 more
The Web Conference 2025, Demo Track
* Equal contribution
Beyond Research
Outside of research, I spend a lot of time around music, especially R&B, jazz, and hip-hop, and I've been playing basketball for years.
I'm drawn to rhythm, taste, and structure, which is probably also how I think about building technology.
“You gotta find what you love.”