Curriculum Vitae

Download the one-page PDF version here

Education

• B.S., Physics & AI in Massachusetts Institute of Technology (Cambridge, MA, United States), Aug. 2024 – Jun. 2027 (Expected)
  • Major: Physics (Course 8), Artificial Intelligence and Decision Making (Course 6-4)
  • GPA: 5.00/5.00 (as of Jun. 2025)
  • Selected Coursework: Machine Learning, Deep Learning, Natural Language Processing, Design and Analysis of Algorithms, Computational Architecture, Probability and Statistics, Quantum Physics I, II & III, Physics of Solids, Statistical Mechanics II, Quantum Field Theory I
• B.S., Physics in Peking University (Beijing, China), Sep. 2023 – Jul. 2024
  • Major: Physics
  • GPA: 3.91/4.00
  • Selected Coursework: Introduction to Computation, Data Structure and Algorithms, Classical Mechanics, Quantum Mechanics, Thermodynamics and Statistical Mechanics, Quantum Statistical Physics, Lie Groups and Lie Algebras
  • Transfer to MIT after one year of study
• High School in No.1 Middle School Affiliated to Central China Normal University (Wuhan, China), Sep. 2020 – Jul. 2023

Working Experiences

• AI & Material Science Research Assistant @ MIT (Cambridge, MA, United States), Jan. 2025 — Present
  • Extensively leveraging various AI techniques in material and defect engineering, including convolutional neural networks, reinforcement learning methods, generative models, and machine learning force fields (MLFF)
  • Co-authored three papers in machine learning-based defect engineering and generative model-powered valence-informed material discovery
  • Supervisor: Professor Mingda Li
• Large Language Model Data Researcher @ ByteDance (Beijing, China), Aug. 2025 — Sep. 2025
  • Served as the Seed LLM data partner, led the design of benchmarks comprised of 500+ physics problems, from university to PhD level
  • Contributed to ongoing internal projects aimed at improving model reasoning fidelity on symbolic tasks
  • Co-led the development of a test-time scaling (TTS) pipeline that enabled gold-medal performance of AI models on IPhO 2025 theoretical problems
• AI Workload Deployment Intern @ Intel (Shanghai, China), Jun. 2025 — Jul. 2025
  • Researched and authored internal tutorials on GPU memory layouts, including CuTe layout and linear layout
  • Analyzed the CuTeDSL lowering process to backend kernels using GEMM as a case study, informing optimizations for AI workload deployment
  • Authored internal technical notes adopted by the CuTeDSL framework team for performance reference

Achievements

• International Physics Olympiad (IPhO) Gold Medalist, Jul. 2023
  • Represented China at the 53rd International Physics Olympiad (one of 5 team members)
  • Ranked 1st place in the world and awarded a Gold Medal
• Chinese Physics Olympiad (CPhO) Gold Medalist, Dec. 2021 & Oct. 2022
  • Awarded the Gold Medal in the finals of the 39th Chinese Physics Olympiad and selected for the national team.
  • Awarded the Gold Medal in the finals of the 38th Chinese Physics Olympiad.

Publications

2025

1. A Foundation Model for Non-Destructive Defect Identification from Vibrational Spectra
Mouyang Cheng^†,*, Chu-Liang Fu^†, Bowen Yu^†, Eunbi Rha, Abhijatmedhi Chotrattanapituk, Douglas L Abernathy, Yongqiang Cheng, and Mingda Li^*
†These authors contributed equally. ^*Corresponding author.

• In review, paper available at: arXiv:2506.00725

2. Enhancing Materials Discovery with Valence Constrained Design in Generative Modeling
Mouyang Cheng^†,*, Weiliang Luo^†, Hao Tang^†, Bowen Yu, Yongqiang Cheng, Weiwei Xie, Ju Li, Heather J. Kulik, and Mingda Li^*
†These authors contributed equally. ^*Corresponding author.

• In review, paper available at: arXiv:2507.19799

3. Reinforcement learning-guided optimization of critical current in high-temperature superconductors
Mouyang Cheng^†,*, Qiwei Wan^†, Bowen Yu^†, Eunbi Rha, Michael J. Landry, and Mingda Li^*
†These authors contributed equally. ^*Corresponding author.

• In review, paper available at: arXiv:2510.22424

Projects

• PaperPlay (HackMIT 2025), Sep. 2025
  • Working in a group of four, developed a web platform that converts hand-drawn Mario-style levels into playable platformers in less than 5 minutes by leveraging OpenCV-based image recognition and deployed via Modal
  • Won 2nd place in the Modal sponsor track
  • Demo available at: https://demo-description.vercel.app/; Code available at: https://github.com/HACKMIT-2025
• Representation Efficiency in Neural Reasoning (MIT 6.7960 Deep Learning), Oct. 2025 — Dec. 2025
  • In a team of three, designed and executed a multilingual evaluation of mathematical reasoning on MMATH and GSM8K, comparing frontier APIs (ChatGPT-5.1, Gemini-2.5-Flash, DeepSeek-V3.2) and open-source 8B models (Qwen3-8B, Llama-3.1-8B-Instruct) across English, Chinese, Spanish, and Thai
  • Performed detailed token-length and character-length analyses under five tokenizers to quantify an Encoder Gap between intrinsic (character-level) and realized (token-level) representation density, showing up to 5-10% token savings in well-aligned multilingual setups
  • Implemented LoRA fine-tuning of Llama-3.1-8B on Chinese GSM8K, improving Chinese accuracy by 8.8 percentage points without degrading English performance, demonstrating that representation bottlenecks are largely learnable rather than architectural
  • Project repo available at: https://github.com/bowenyu066/language-shapes-reasoning/; blog post available at: https://bowenyu066.github.io/posts/notes/multilingual-math-reasoning/

Skills

• Languages: English (fluent), Chinese (native), Cantonese (basic), Spanish (basic)
• Programming Languages: C/C++, Python, Bluespec, RISC-V Assembly, LaTeX
• Libraries: PyTorch, NumPy, Matplotlib, ASE, PyMatgen, Matformer, MACE, Triton, CUTLASS

Hobbies

• Long-distance running: As of Dec. 2025, I have completed 1 half marathon, with the following personal bests:
  • Half marathon: 1:48:56 (2023 Wuhan Marathon)
• Table tennis
• Movies
• More to explore…

---

Last updated: December 12, 2025