Westlake University
B.S. in Artificial Intelligence
- 2025–26 GPA: 3.787; full scholarship recipient.
- Coursework: data structures & algorithms, linear algebra, calculus, physics, AI/CS.
Biography & experience
A student of artificial intelligence, interested in the engineering that makes intelligent systems small, fast, and dependable.
At Westlake University, I study artificial intelligence with a focus on efficiency. I am drawn to questions that sit between machine learning and computer systems: how a model can fit inside a microcontroller, how an agent can reason about low-level code, and how careful engineering can turn a promising idea into something that runs reliably. My current explorations include efficient sequence models for keyword spotting and agentic approaches to MCU kernel optimization. This site is a record of that work and a place for writing that does not fit in a paper.
This site collects research notes, longer project write-ups, and Between the Lines, a space for ideas that do not fit into a paper. It is maintained as a static website: the writing lives in plain text, and the design stays out of the way.
B.S. in Artificial Intelligence
ENCODE Lab · Westlake University
Supervisor: Prof. Huan Wang
Investigating hardware-aware neural networks on ARM Cortex-M microcontrollers, including architecture, quantization, C/C++ implementation, memory profiling, and host/device verification. Developing reproducible experimentation pipelines that combine structured agent outputs, compiler/build gates, static and semantic checks, automated tests, and MCU measurements.
Westlake Chemistry 2026 Summer Undergraduate Research Internship · Westlake University
Prof. Yiqin Gao's group · Computational chemistry. Conducted SPONGE molecular-dynamics experiments on 862 TIP3P water molecules and analyzed the oxygen–oxygen radial distribution function to characterize liquid-water structure.
Research project
Exploring state-space models for always-on keyword spotting within the memory and latency limits of microcontrollers.
Research project
Investigating whether language-model agents can help write, test, and refine the low-level kernels that make inference practical on microcontrollers.
National level
The TinyMamba project was selected at the national level for undergraduate innovation and entrepreneurship training.
Finalist in the university-level innovation competition with the TinyMamba project.
Technology track
Selected for the AI and technology track of the Ebhar Misr program.
Supported high-school students with Su–Schrieffer–Heeger (SSH) circuit experiments, oscilloscope measurements, and troubleshooting.
Zhumadian
Led five “Tiny AI” sessions for local middle- and high-school students, translating audio recognition and embedded-AI research into interactive lessons.
Beyond research, I keep a space for writing that does not fit into a paper: philosophy, psychology, feelings, observations, questions, and stories. It is called Between the Lines ✨.