Ze-Wei Liou

PhD student @ Princeton

Princeton University

zewei.liou [at] princeton [dot] edu

About Me

📢 Check out my new blog post! ReplaySSM

I am a first-year PhD student at Princeton, advised by Prof. Tri Dao. My research focuses on ML Systems.

My official name is Ze-Wei Liou, but I also go by the name “Johnny.” I did my undergrad in EE at National Taiwan University. Feel free to reach out!

Education

PhD ECE

2025-09 ~

Princeton University

BS EE

2020-09
2024-06

National Taiwan University

📚 My Research

I am broadly interested in efficient computing for ML. Currently, my research focuses on ML systems, specifically optimizing inference serving for recurrent state space models (Mamba-2, Gated DeltaNet).

Past Experience

Previously, I worked on scheduling and mapping for AI accelerators and in-memory computing, topics I remain excited about and happy to discuss. I interned at IBM Research (San Jose) under Dr. Geoffrey Burr and Dr. Sidney Tsai. I also worked as a research assistant at Academia Sinica with Dr. Ding-Yong Hong and at NTU with Prof. An-Yeu Wu.

Publications

2026

Y.-C. Chen, C. P. Lee, Ze-Wei Liou, N. Verma (2026). Massive Spikes in LLMs are Bias Vectors: Mechanistic Uncovering and Spike-Free Quantization. arXiv ‘26.

Preprint

2025

Ze-Wei Liou, D.-Y. Hong (2025). Optimizing Compute Core Assignment for Dynamic Batch Inference in AI Inference Accelerator. ACM SAC ‘25.

DOI

2023

G. W. Burr, H. Tsai, W. Simon, I. Boybat, S. Ambrogio, C.-E. Ho, Ze-Wei Liou, et al. (2023). Design of Analog-AI Hardware Accelerators for Transformer-based Language Models. IEDM ‘23.

DOI