David Zhao

David Zhao

I am a 5th year Ph.D. student in the joint program in Statistics and Machine Learning at Carnegie Mellon University. My research focuses on developing new methods to tackle challenging real-world problems in the physical sciences. In particular, I work on uncertainty quantification and likelihood-free inference, with applications to astronomy and physics. I am fortunate to be advised by Professor Ann B. Lee.

Prior to CMU, I obtained an M.Sc. in Statistics from ETH Zurich and an A.B. in Applied Mathematics from Harvard University. I also have 3+ years of work experience in the quantitative finance industry, in both investment research and proprietary trading.

  • Statistical Methodology
  • Machine Learning
  • Applied Quantitative Research
  • Ph.D. in Statistics and Machine Learning, 2023

    Carnegie Mellon University

  • M.Sc. in Statistics, 2016

    ETH Zurich

  • A.B. in Applied Mathematics, 2013

    Harvard University

Selected Papers and Talks

Last updated Jan 2022

(+) denotes Equal Contribution

(2021). Diagnostics for Conditional Density Models and Bayesian Inference Algorithms. 37th Conference on Uncertainty in Artificial Intelligence (UAI 2021). Spotlight Talk.

Preprint PDF Code

(2021). MD-split+: Practical Local Conformal Inference in High Dimensions. 38th International Conference on Machine Learning (ICML 2021). Workshop on Distribution-Free Uncertainty Quantification.

Preprint Code

(2021). Invited Talk 04/29/2021: ''Validating Conditional Density Models and Posterior Estimates''. Informatics Statistics and Science Collaboration (ISSC) of the Rubin Observatory Legacy Survey of Space and Time (LSST).

Code Slides Video

(2019). Cryptocurrency Price Prediction and Trading Strategies Using Support Vector Machines. arXiv preprint.


Work Experience

Quantitative Research Intern
Jun 2022 – Aug 2022 New York, NY
  • Alpha research for equity feature forecasting
  • Returning full-time in Summer 2023
Quantitative Trading Intern
Jan 2021 – Jun 2021 New York, NY
  • Developed technical indicators for predicting 30 min and close-to-close single stock returns. Incorporated order book history, Barra factors, GICS, and short interest information.
  • Combined indicators into tradable model using tree boosting (e.g. CatBoost, LightGBM) methods. Model added value and had low correlation to team’s existing strategies.
Derivatives Trader
Mar 2017 – Mar 2018 Chicago, IL
  • Oil options trader on a small team with shared responsibility for PnL and risk every day.
  • Made markets and trading decisions in real time, adjusted models to changing market conditions, actively monitored opportunities and risks, handled day-to-day operations.
Research Analyst
Jul 2013 – May 2015 Greenwich, CT
  • Researcher on product providing country and currency strategy exposure without derivatives.
  • Researcher on optimizing account rebalance schedules to reduce turnover and market impact.
  • Researcher on tax loss harvesting strategy that realizes capital gains on a tax-favorable schedule.


  • davidzhao AT stat DOT cmu DOT edu