Chaoqi Luo
Chaoqi Luo

Chaoqi Luo

罗朝旗

M.S. student, EE (AI Track)
University of Pennsylvania

luo31@seas.upenn.edu

I am Chaoqi Luo, an M.S. student in Electrical Engineering (AI track) at the University of Pennsylvania. I am currently a Research Engineer Intern at Zhipu (Z.AI), where my work focuses on agent harness engineering—building the runtime, tools, context-management systems, and feedback loops that make LLM agents more capable and reliable in real-world settings. At Penn, I also work as a research intern in Prof. Jiatao Gu’s group, collaborating with Yong-Hyun Park (PhD @ UPenn CIS) on accelerating autoregressive diffusion video generation with few-step flow distillation.

I did my undergraduate study at the SWJTU–Leeds Joint School (Chengdu, China), majoring in Electronic & Electrical Engineering and graduating with UK First-Class Honours. I was fortunate to do research advised by Prof. Zitong Yu.

Work Experience

  • Zhipu AI (Z.AI)

    Research Engineer Intern, AI Agents

    2026.07 – Present

    Beijing, China

Education

  • University of Pennsylvania

    M.S. in Electrical Engineering · AI track

    2025.08 – Present

    Philadelphia, US

  • Southwest Jiaotong University

    B.Eng. in Electronic Information Engineering

    2021.09 – 2025.07

    Chengdu, China

  • University of Leeds

    B.Eng. in Electronic & Electrical Engineering · UK First-Class Honours

    2021.09 – 2025.07

    Leeds, UK

Selected Publications

View all
  1. Pi-E-Flow: Uncertainty-Guided Flow Distillation for Autoregressive Video Generation

    Y-H Park, H Chen, C Luo, K Yu, J Gu

    SPIGM @ ICML2026

  2. PhysMamba: Efficient remote physiological measurement with SlowFast temporal difference Mamba

    C Luo, Y Xie, Z Yu

    Chinese Conference on Biometric Recognition (CCBR), 2024 · ★ Best Paper Honorable Mention

  3. Semi-rPPG: Semi-supervised remote physiological measurement with curriculum pseudo-labeling

    B Wu, Z Yu, Y Xie, W Liu, C Luo, Y Liu, RSM Goh

    IEEE Transactions on Instrumentation and Measurement (TIM), 2025

Writing

All