David X. Wu

I am a Ph.D. student at UC Berkeley, where I’m extremely fortunate to be coadvised by Prasad Raghavendra and Anant Sahai. I’m broadly interested in problems related to theoretical computer science and statistics, such as Markov chains, statistical inference, and machine learning theory. Recently, I have also been interested in LLM reasoning and RL for coding and math. Last summer, I was an ML research intern at Windsurf working on coding agent capabilities and evals. I’m grateful to be supported by an NSF GRFP fellowship and an OpenAI Superalignment Grant.

I graduated with a B.Sc. from MIT with a double major in mathematics (Course 18) and computer science (Course 6-3), where I was lucky to do research with Justin Solomon and Suvrit Sra. As an undergrad, I interned in quant finance at HRT and Akuna Capital.

Selected publications

STOC

Weak Poincaré Inequalities, Simulated Annealing, and Sampling from Spherical Spin Glasses

Brice Huang, Sidhanth Mohanty, Amit Rajaraman, and David X Wu

ACM Symposium on Theory of Computing, 2025

arXiv Bib

@article{huang2024weak,
  title = {Weak Poincar\'{e} Inequalities, Simulated Annealing, and Sampling from Spherical Spin Glasses},
  author = {Huang, Brice and Mohanty, Sidhanth and Rajaraman, Amit and Wu, David X},
  journal = {ACM Symposium on Theory of Computing},
  year = {2025}
}

ICLR

Provable Weak-to-Strong Generalization via Benign Overfitting

David X Wu, and Anant Sahai

International Conference on Learning Representations (Preliminary version at NeurIPS’24 M3L Workshop), 2025

arXiv Bib

@article{wu2024provable,
  title = {Provable Weak-to-Strong Generalization via Benign Overfitting},
  author = {Wu, David X and Sahai, Anant},
  journal = {International Conference on Learning Representations (Preliminary version at NeurIPS'24 M3L Workshop)},
  year = {2025}
}

FOCS

Locally Stationary Distributions: A Framework for Analyzing Slow-Mixing Markov Chains

Kuikui Liu^*, Sidhanth Mohanty^*, Prasad Raghavendra^*, Amit Rajaraman^*, and David X Wu^*

IEEE Annual Symposium on Foundations of Computer Science, 2024

arXiv Bib

@article{liu2024locally,
  title = {Locally Stationary Distributions: A Framework for Analyzing Slow-Mixing Markov Chains},
  author = {Liu, Kuikui and Mohanty, Sidhanth and Raghavendra, Prasad and Rajaraman, Amit and Wu, David X},
  journal = {IEEE Annual Symposium on Foundations of Computer Science},
  equal_contribution = {Liu, Kuikui and Mohanty, Sidhanth and Raghavendra, Prasad and Rajaraman, Amit and Wu, David X},
  year = {2024}
}

FOCS

Fast Mixing in Sparse Random Ising Models

Kuikui Liu^*, Sidhanth Mohanty^*, Amit Rajaraman^*, and David X Wu^*

IEEE Annual Symposium on Foundations of Computer Science, 2024

arXiv Bib

@article{liu2024fast,
  title = {Fast Mixing in Sparse Random Ising Models},
  author = {Liu, Kuikui and Mohanty, Sidhanth and Rajaraman, Amit and Wu, David X},
  journal = {IEEE Annual Symposium on Foundations of Computer Science},
  equal_contribution = {Liu, Kuikui and Mohanty, Sidhanth and Rajaraman, Amit and Wu, David X},
  year = {2024}
}

STOC

Robust recovery for stochastic block models, simplified and generalized

Sidhanth Mohanty^*, Prasad Raghavendra^*, and David X Wu^*

ACM Symposium on Theory of Computing, 2024

arXiv Bib

@article{mohanty2024robust,
  title = {Robust recovery for stochastic block models, simplified and generalized},
  author = {Mohanty, Sidhanth and Raghavendra, Prasad and Wu, David X},
  journal = {ACM Symposium on Theory of Computing},
  year = {2024},
  equal_contribution = {Mohanty, Sidhanth and Raghavendra, Prasad and Wu, David X}
}