I am a fourth year Electrical Engineering PhD student at Stanford University advised by Professor John Duchi and supported by the 3-year Professor Michael J. Flynn SGF fellowship. I have been fortunate to work with Moritz Hardt this past summer in the intersection of causal inference and control theory. My research interests are broadly in causal inference, optimization, and differential privacy, and I am currently working on projects in causal inference and adaptive optimization.

Prior to joining Stanford, I was fortunate as an undergraduate at UC Berkeley to work with Professors Jean Walrand, Laurent El Ghaoui, and Kannan Ramchandran. I was also fortunate to be a teaching assistant for Data Structures (CS61b) in Sp'17, Algorithms (CS170) in Fa'17, and Probability (EE126) in Sp'18 and Sp'19. In 2019, I was awarded the UC Berkeley Campus Outstanding GSI award.


Gary Cheng*, Hilal Asi*, Karan Chadha*, and John Duchi. "Private optimization in the interpolation regime: faster rates and hardness results." Spotlight Presentation at ICML 2022.

Gary Cheng*, Karan Chadha*, and John Duchi. "Accelerated, Optimal, and Parallel: Some Results on Model-Based Stochastic Optimization." Spotlight Presentation at ICML 2022.

Gary Cheng, Zachary Charles, Zachary Garrett, and Keith Rush. "Does Federated Dropout actually work?" Presentation at CVPR FedVision 2022.

Gary Cheng*, Karan Chadha*, and John Duchi. "Fine-tuning in Federated Learning: A simple but tough-to-beat baseline." arXiv preprint.

Tavor Baharav, Gary Cheng, Mert Pilanci, David Tse. "Approximate Function Evaluation via Multi-Armed Bandits." Poster at AISTATS 2022.

Gary Cheng*, Hilal Asi*, Karan Chadha*, and John Duchi. "Minibatch Stochastic Approximate Proximal Point Methods." Spotlight Presentation at Neurips 2020 (video recording).

Gary Cheng, Kabir Chandrasekher, and Jean Walrand. "Static and Dynamic Appointment Scheduling with Stochastic Gradient Descent." In American Control Conference 2019.

Gary Cheng, Armin Askari, Kannan Ramchandran, and Laurent El Ghaoui. "Greedy Frank-Wolfe Algorithm for Exemplar Selection." Poster at BayLearn 2018.

* denotes equal contribution

If you would like contact me, please email me at


image from xkcd.com