Portrait of Haoan Feng

Ph.D. Candidate in Computer Science
University of Maryland, College Park

Haoan Feng

Spatial intelligence through world models built on explicit 3D representations: neural fields, Gaussian splats, and triangulated meshes, where scene structure is something you can query, differentiate, and edit.

Open to full-time Research Scientist / Engineer roles in 3D visionGraduating Summer 2027

FFGaussType Text and icons become relightable Gaussian assets inside 4D-Gaussian video, in one 9.1 ms forward pass.

arXiv 2026 · Dolby ATG internshipPaper & links

Papers at

  • CVPR 2026
  • ACM SIGSPATIAL 2026
  • ACM TSAS

Research internships

  • Dolby Laboratories ATG (2025, 2026)

Recognition

  • SIGSPATIAL ’24 Best Paper Runner-up
  • OASIS ’26 Top-10 Finalist

Reviewer

  • CVPR
  • ICCV
  • ECCV
  • NeurIPS

Selected work

Research

All publications

Biography

I am a Ph.D. candidate in Computer Science at the University of Maryland, College Park, advised by Prof. Leila De Floriani. My research asks how machines should represent the 3D world so they can measure it, reason about it, and edit it. I look for representations that are continuous like the world itself, compact enough to scale, and structured enough to stay interpretable.

I have pursued this question across very different kinds of data: neural fields and topology-aware analysis for terrain and scientific data and, during two internships at Dolby Laboratories’ Advanced Technology Group, geometry for video-language models and native 3D assets for volumetric video. The common thread is turning raw samples into structure that downstream systems can use.

Before UMD, I studied 3D vision and graphics with Prof. Long Quan at HKUST (BEng, MPhil). I am drawn to work where frontier research turns into real products, and I am looking for full-time research scientist or engineer roles starting in Summer 2027.

Interests

  • Spatial Intelligence
  • World Models with Explicit 3D Representations
  • Spatial Representation Learning
  • Neural Rendering
  • Topological/Morphological Analysis
  • May 2026 – Aug 2026

    PhD Research Intern

    Dolby Laboratories – Advanced Technology Group (ATG)

  • Jun 2025 – Aug 2025

    PhD Research Intern

    Dolby Laboratories – Advanced Technology Group (ATG)

  • 2021 – 2027

    PhD in Computer Science

    University of Maryland, College Park

  • 2018 – 2020

    MPhil in Computer Science

    Hong Kong University of Science and Technology

  • 2014 – 2018

    BEng Dual Major in Computer Science Engineering and Electronic and Computer Engineering

    Hong Kong University of Science and Technology

“Topology is precisely the mathematical discipline that allows the passage from local to global.” — René Thom

  1. Nov. 2026
    New preprint FFGaussType: Feed-Forward Conversion of Glyph and SVG Geometry into Compact Gaussian Assets, from my 2026 internship at Dolby Laboratories – Advanced Technology Group (ATG). It brings text and icons into volumetric videos built on 4D Gaussians: one 9.1 ms forward pass turns a glyph or SVG outline into a relightable Gaussian asset. Paper on arXiv, videos on the project page.
  2. Nov. 2026
    At ACM SIGSPATIAL in Riverside, I presented ImplicitTerrainV2 in an oral session and our work on amortized terrain representations in a lightning talk and poster session. Xin Xu, Yutian Tao, and I also participated in the OASIS Student Challenge as a team with our urban flood screening project.
  3. Oct. 2026
    The preprint for our topology-guided GeoAI agent is now available. The project is a top-10 finalist at OASIS 2026, the ACM SIGSPATIAL student challenge. Our team received a Google DeepMind Student Travel Grant and was selected for an on-site presentation and demo.
  4. Aug. 2026
    ImplicitTerrainV2 (full paper) and Amortized Terrain (poster) accepted at ACM SIGSPATIAL 2026.

Outside the lab

Things I build for fun

All projects

Meowverse

iPhone & iPad app · Apple Vision Pro in development

Let your photos grow into your cat’s little world.

Shipped to TestFlight · Swift 6 · SwiftUI · RealityKit · 344 core-library tests

Bella the cat, asleep on a blue blanket

Bella, my cat, and the muse behind Meowverse.

  • Hiking
  • Cooking
  • Reading & Manga
  • 3D printing