I am Tao Hu, an undergraduate researcher driven by curiosity for mathematical optimization, stochastic control, and learning theory. Currently studying mathematics at Xi’an Jiaotong University and as an intern at Cardinal Operation, I explore how elegant theory powers practical algorithms.

Get in touch See my work
Portrait of Tao Hu

Welcome!

I am a mathematics major in the Honor Science Program at Xi’an Jiaotong University, currently ranking first in my cohort with a 4.19/4.3 GPA. In Spring 2025, I joined the Berkeley Global Access Program at UC Berkeley to dive deeper into graduate-level optimization and stochastic processes.

My research sits at the intersection of optimization theory and stochastic control. I enjoy formulating elegant algorithms, experimenting with scientific computing tools, and sharing knowledge through teaching and mentorship.

Current Interests

  • Physics-informed neural networks and scientific machine learning
  • Stochastic and derivative-free optimization
  • Numerical methods for differential equations
  • Formalizing mathematics in proof assistants

Education

Xi’an Jiaotong University

Honor Science Program in Mathematics • Sept. 2022 – Present

GPA: 4.19/4.3 (Rank 1/32). Completed advanced coursework in real and complex analysis, probability theory, topology, abstract algebra, numerical analysis, and differential equations.

University of California, Berkeley

Berkeley Global Access • Jan. 2025 – May. 2025

Graduate coursework in Mathematical Programming, Applied Stochastic Process, Stochastic Optimization, and Numerical Solution of Differential Equations. GPA: 4.0/4.0.

Xi’an Jiaotong University

Honor Youth Program (Special Class for the Gifted Young) • Sept. 2021 – July. 2022

Ranked 1/40 in a rigorous accelerated curriculum designed for exceptional young scholars.

Research Experience

Research Assistant • Cardinal Operations

Shanghai, China • Apr. 2025 – Present

  • Investigating strategies to improve physics-informed neural network training with DRSOM and MUON optimizers.
  • Exploring Polyak step sizes for training scientific machine learning models.
  • Proposed a distributionally robust steepest descent method for resilient optimization.
  • Built an extensible benchmarking repository for comparing optimizers on nanoGPT and CNN workloads.

Research Assistant • University of California Berkeley

Berkeley, California, USA • Mar. 2025 – May. 2025

  • Finished a final project in the graduate course "Mathematical Programming II" about the performance of Frank-Wolfe for nonconvex functions under inexact gradient and computational comparison between linear minimization oracle and proximal operators.
  • Improved the result of a paper which was later presented at the ICCOPT 2025 Conference in the final project.
  • Established the convergence guarantee of Frank-Wolfe under non-convex setting with inexact gradient.

Participant • Mathematics in Lean Workshop

Department of Mathematics, Peking University • Jan. 2025

  • Learned to formalize optimization proofs using the Lean theorem prover, strengthening my ability to connect rigorous mathematics with modern verification tools.
  • Used Lean to prove an equivalent characterization of intrinsic interior.

Academic & Volunteer Experience

Teaching Assistant / Grader

NYU Shanghai • Jan. 2026 – Present

Courses:
  • Network and Dynamics & Honors Analysis I
  • Responsible for grading assignments and providing detailed feedback on student coursework.

Teaching Assistant — Higher Algebra

Department of Mathematics, Xi’an Jiaotong University • Sept. 2024 – Jan. 2025

  • Designed and led bi-weekly tutorials for a class of over 50 undergraduate students.
  • Created comprehensive solution guides and graded assignments to reinforce conceptual understanding.

Mathematics Tutor (Volunteer)

School of Journalism and New Media, Xi’an Jiaotong University • Jan. 2024 – Feb. 2024

  • Mentored a 10th-grade student in mathematics, focusing on problem-solving strategies.
  • Facilitated significant academic improvement, helping the student advance from rank 11,594 to 3,870 in city-wide examinations.

Honors & Awards

Top Ten Outstanding Student Models

Xi’an Jiaotong University • 2023 – 2024

Highest undergraduate honor at XJTU (Rank 1/10).

National Scholarship

2022 – 2023

Awarded to the top 0.25% of students nationwide.

Baosteel Excellent Student Award

2023 – 2024

National recognition for academic excellence, one of 500 recipients nationwide.

Professor Jianhua Wang Scholarship

2024 – 2025

Awarded to top graduating undergraduates at Xi’an Jiaotong University for outstanding academic performance, integrity, and leadership.

Chinese Mathematics Competition

National 1st Prize • 2022 & 2023

Recognized for excellence in mathematical problem solving.

Chinese Mathematics Competition (Final Round)

Senior Group • 2024

1st Prize in the national finals.

Skills

Programming

  • Python
  • MATLAB
  • C / C++ / C#
  • Lean
  • LaTeX

Languages

  • English — TOEFL 105
  • GRE — 322 (V+Q) & 3.0 AWA

Subject Mastery

  • GRE Mathematics Subject Test: 970
  • Advanced undergraduate analysis and algebra
  • Stochastic processes & optimization

Blogs

Exploring the Robustness of the Frank-Wolfe Method

March 2025  |  Optimization

A deep dive into Frank-Wolfe algorithms under inexact gradient information and a comparison between linear minimization oracles and projections.

Read the full article

A Wasserstein Penalty Framework for Stochastic Optimization

August 2025  |  Optimization

Interpreting stochastic gradient updates through distributionally robust penalties, including explicit algorithms and full convergence proofs.

Read the blog post

Training Physics Informed Neural Network with MUON

August 2025  |  Optimization

Optimizing Physics Informed Neural Network with MUON

Read the blog post

First Attempt to Reproduce a State-of-the-Art Article

August 2024  |  Machine Learning

Learned to Use Pytorch and Data-Gathering Techniques

Read the blog post

Let’s Connect

I am always excited to discuss optimization, mathematics, and opportunities for collaboration. Feel free to reach out!

No. 28, West Xianning Road, Xi’an 710049, Shaanxi, China