vweng.org

Haodi Weng

I’m a PhD student at The Hong Kong Polytechnic University. My research lies in 3D scene understanding and autonomous driving.

I currently study how to understand indoor scenes from sparse views. I also work on end-to-end driving planners, with a focus on handling long-tail scenarios and reducing the computation needed at inference time.

Outside the lab, I enjoy travel photography.

Education

  1. Now
    PhD studentThe Hong Kong Polytechnic University
  2. 2024
    MSc in Computational Statistics and Machine LearningUniversity College London✦   Dean’s List · Top 5%
  3. 2023
    MRes in Artificial Intelligence and Machine LearningImperial College London
  4. 2022
    BSc in Computer ScienceUniversity College London

Publications

IEEE ISMAR · 2025

PersoNo: Personalised Notification Urgency Classifier in Mixed Reality

Jingyao Zheng*, Haodi Weng*, Xian Wang, Chengbin Cui, Sven Mayer, Chi-lok Tai, and Lik-Hang Lee

* Equal contribution

PDFDOI

About this paper

Which notifications are worth interrupting you for when you’re wearing a mixed reality headset? We studied how the answer depends on what you’re doing and built a classifier that learns individual preferences.

Patents

U.S. patent application · Published 12 March 2026

Generating Multi-Task Frame Predictions

Haodi Weng, Felix John Samuel Bragman, Danail V. Stoyanov, and Imanol LuengoMuntion

US 2026/0073189 A1

Career

Jun 2026 – present

Research Intern

千里智行 · Afari Technology

I work on world models for rare driving scenarios and led the development of a looped Transformer for end-to-end driving.

Aug – Dec 2025

Research Assistant

The Hong Kong Polytechnic University

I studied how models answer questions about 3D scenes, from representing a scene at multiple scales to interpreting the model’s answers.

Jun – Sep 2024

AI/ML Engineering Intern

Medtronic

I developed a shared model for recognising surgical phases, segmenting video frames and detecting instruments in surgical footage.