Cheng Zeng

Postdoctoral associate at University of Florida.

prof_pic.jpg

Phone: +1 401-396-6668

Email: c.zeng@ufl.edu

Hi! I’m currently a postdoctoral researcher at the University of Florida. Before joining UF, I was a postdoctoral fellow at the Institute of Experiential AI and Roux Institute of Northeastern University. I received my Ph.D. in Chemical Engineering and an M.S. in Data Science in 2022 from Brown University.

My research lies at the intersection of AI, Materials Science and Chemistry. I develop and deploy generative models, machine learning interatomic potentials, and phenomenological models to accelerate materials and molecular design for a sustainable future.

I’m currently on the job market—welcome to connect via LinkedIn!

Recent news [archive]

2026/03 MolCrystalFlow data is available at zenodo. Code is available at Github.
2026/02 Excited to introduce MolCrystalFlow, now available at arXiv. We leverage a hierarchical representation combined with an equivariant Riemannian flow matching to generate crystal packing for given rigid-body molecular conformers. We benchmark MolCrystalFlow using two open-source datasets and against state-of-the-art generative model for large-size periodic crystals and rule-based Genarris-3 generation method. Code and data to follow soon…
2026/01 Our PropMolFlow work is finally online at Nature Computational Science. News release is here.

Selected publications [full list]

  1. molcrystalflow.png
    MolCrystalFlow: Molecular Crystal Structure Prediction via Flow Matching
    Cheng Zeng, Harry W. Sullivan, Thomas Egg, Maya M. Martirossyan, Philipp Höllmer, Jirui Jin, Richard G. Hennig, Adrian Roitberg, Stefano Martiniani, Ellad B. Tadmor, and Mingjie Liu
    Feb 2026
    arXiv:2602.16020 [cs]
  2. molguidance.png
    MolGuidance: Advanced Guidance Strategies for Conditional Molecular Generation with Flow Matching
    Jirui Jin, Cheng Zeng, Pawan Prakash, Ellad B. Tadmor, Adrian Roitberg, Richard G. Hennig, Stefano Martiniani, and Mingjie Liu
    Dec 2025
    arXiv:2512.12198 [cs]
  3. propmolflow.png
    PropMolFlow: property-guided molecule generation with geometry-complete flow matching
    Cheng Zeng, Jirui Jin, Connor Ambrose, George Karypis, Mark Transtrum, Ellad B. Tadmor, Richard G. Hennig, Adrian Roitberg, Stefano Martiniani, and Mingjie Liu
    Nature Computational Science, Jan 2026
  4. d_vae_hea.png
    Data-efficient and Interpretable Inverse Materials Design using a Disentangled Variational Autoencoder
    Cheng Zeng, Zulqarnain Khan, and Nathan L. Post
    AI & Materials, Dec 2024
  5. ml-hea-corr.png
    Machine learning accelerated discovery of corrosion-resistant high-entropy alloys
    Cheng Zeng, Andrew Neils, Jack Lesko, and Nathan Post
    Computational Materials Science, Mar 2024
  6. nft.png
    A nearsighted force-training approach to systematically generate training data for the machine learning of large atomic structures
    Cheng Zeng, Xi Chen, and Andrew A. Peterson
    The Journal of Chemical Physics, Feb 2022