Hello, I’m Zhengxu! 👋

I’m a Computer Science M.S. student at Stanford University and a UC Berkeley graduate in Computer Science and Data Science with High Distinction. My current work focuses on scalable model improvement, structured feedback, verifiable code generation, and empirical analysis of datasets and claims in NLP research.

Academic Interests 🎓

My main interests sit at the intersection of machine learning systems and applied research. I’m especially interested in:

  • Large language models and agentic systems
  • Model feedback, critique, and iterative refinement
  • Verifiable code generation and evaluation
  • Dataset intelligence and scientific knowledge extraction
  • Deep Reinforcement Learning
  • Systems Programming

Across Stanford and Berkeley, I’ve worked on projects ranging from benchmarking verifiable code generation and serving diffusion transformers to building reinforcement learning systems for healthcare operations and metadata pipelines for NLP research.

Professional Journey 💻

I’m currently a Graduate Research Assistant at Stanford’s IRIS Lab and SALT Lab, where I work on feedback-driven model improvement and large-scale extraction of structured claims from NLP papers. My recent experience also includes:

  • Research at UC Berkeley’s Sky Computing Lab on efficient serving systems for diffusion transformers
  • Research at UC Berkeley’s College of Environmental Design on reinforcement learning for smart hospital operations
  • Machine learning engineering at UC Berkeley School of Public Health on infectious disease preprint triage and review tooling
  • Teaching support for Berkeley’s Large Language Model Agents course
  • Applied NLP research with the University of Nebraska and backend engineering in startup environments

Beyond the Code 🎮

When I’m not coding or working on projects, you can find me:

  • Playing soccer with friends
  • Competing in League of Legends (always striving to improve my gameplay!)
  • Enjoying a game of badminton
  • Staying updated with the latest tech developments and research papers

Let’s Connect!

I’m always interested in connecting with researchers, builders, and potential collaborators. Feel free to reach out if you’d like to talk about ML systems, research ideas, or opportunities at the intersection of machine learning and real-world impact.

“The best way to predict the future is to invent it.” - Alan Kay