I am a postdoc at Princeton University, advised by Ryan Adams and Adji Bousso Dieng, working on artificial intelligence for design and modeling of dynamic processes and materials. I completed my PhD at Carnegie Mellon University, advised by John Kitchin. While there, my research focused on machine learning to overcome challenges of molecular simulation, including fast, accurate models to approximate quantum chemical methods, investigation of physical phenomena by analyzing large datasets of atomic configurations, and uncertainty quantification for models and physical properties. I have also published papers on machine learning for finance and macroeconomics. My interests include machine learning, atomic simulations, blockchain, and finance.

While at CMU, I also received an MS in machine learning and interned at JPMorgan Chase, working on strategy at the Chief Investment Office. Previously, I worked as a chemical engineer and graduated magna cum laude with a BS in chemical engineering from UT Austin.


  • Model-Specific to Model-General Uncertainty for Physical Properties [doi] [github] [PDF]
    Ni Zhan, John R. Kitchin
    Industrial & Engineering Chemistry Research 2021
  • Machine Learning Models and Uncertainty for Atomic Simulations [PDF]
    PhD Dissertation 2021
    Thesis Committee: John Kitchin, Zachary Ulissi, Michael Widom, Aditya Khair, Erik Ydstie
  • Origin of the Stokes-Einstein Deviation in Liquid Al-Si [doi] [code]
    Ni Zhan, John R. Kitchin
    Molecular Simulation 2021
    Presented talk at American Chemical Society Spring Meeting 2021
  • Uncertainty quantification in machine learning and nonlinear least squares regression models [doi][github]
    Ni Zhan, John R. Kitchin
    Aiche Journal 2021
    Presented two talks at Aiche Annual Meeting 2019 and Machine Learning in Science and Engineering Conference 2019
  • Where does the Stimulus go? Deep Generative Model for Commercial Banking Deposits [arXiv] [poster]
    Ni Zhan
    Neurips Workshop on ML for Economic Policy 2020
  • Graphical models for financial time series and portfolio selection [doi]
    Ni Zhan, Yijia Sun, Aman Jakhar, He Liu
    Proceedings of International Conference on AI in Finance 2020
    Presented two talks at International Conference on AI in Finance, 2020 and Toronto Machine Learning Society, 2021

  • Modeling Superalloys using Machine Learned Potential [poster]
    John Kitchin, Jenny Zhan, Michael Widom, Bojun Feng, Jim Lill, Chris Woodward
    Published as Chapter in PhD Dissertation
    Presented talk at Department of Defense High Performance Computing User Group Meeting, 2019


  • Recommender Systems and Statistical Guarantee for Collaborative Filtering [PDF]
    Ni Zhan, Yilin Yang, Zicheng Cai
    CMU 10716 Advanced ML: Theory and Methods, Spring 2021
    Project summarized statistical guarantees of matrix completion problem, following matrix incoherence property (Candes and Recht 2009) and spikiness and rank measures (Negahban and Wainwright 2012). We additionally summarized the extension of the latter to a graph regularized matrix completion, as shown by Rao et al. 2015.
  • Music Recommender System and Image Dimensionality Reduction [PDF]
    Yilin Yang, Ni Zhan
    CMU 10805 ML with Large Datasets (Distributed Computing), Fall 2020
    Music recommender on million songs using collaborative filtering, Spark on AWS EMR.
  • Graph convolutional neural network for atomic structures [PDF]
    Junwoong Yoon, Ni Zhan, Mingjie Liu
    CMU 10707 Deep Learning, Spring 2019
  • Identifying Duplicate Questions using Siamese LSTM Architecture [PDF]
    Junwoong Yoon, Ni Zhan
    CMU 10701 Graduate Introduction to Machine Learning, Fall 2018