mybian.cn

Experience

Research Experience

2025–2026

Institute of High Energy Physics, Chinese Academy of Sciences

National undergraduate innovation and research training program of the Chinese Academy of Sciences · Individual project supervised by Haiyun Teng

GPU-based streaming data acquisition for neutron spectrometers

Implemented direct memory access (DMA) readout from neutron imaging detectors, concurrent GPU stream processing, and AI image analysis tailored to individual experiments.

Built a GPU streaming acquisition prototype for the Phase II neutron spectrometer upgrade and developed a software framework for data acquisition, image analysis, data-stream output debugging, and runtime monitoring.

2025–2026

Institute of Software, Chinese Academy of Sciences

National undergraduate innovation and research training program of the Chinese Academy of Sciences · Individual project supervised by Junjie Wang

MemoDroid · Dynamic hierarchical memory for LLM-based GUI testing

Designed three tiers of memory for testing graphical interfaces with language models. Episodic memory records interaction traces, reflective memory captures failure patterns and redundant actions, and strategic memory transfers exploration strategies across apps. The agent retrieves relevant memories at runtime and adds them to its prompt.

Implemented MemoDroid as a lightweight plug-in for existing LLM-based GUI testing methods and evaluated it on at least 200 real-world apps across 15 categories.

2026–Present

Shenzhen Bay Laboratory

Visiting Student · Institute of Systems and Physical Biology · Yaoqi Zhou Group

Downstream Task Optimization and Multimodal Extension of RESM

During my research visit to Prof. Yaoqi Zhou’s group, I applied contrastive learning to RESM, the group’s RNA language model, to improve its performance on downstream tasks. I also worked on extending RESM to integrate information from multiple modalities.

Education

2024–2028 (expected)

Xinjiang University

Undergraduate in Intelligence Science and Technology · School of Future Technology · Urumqi

Coursework includes Advanced Mathematics, Linear Algebra, Probability and Mathematical Statistics, Data Structures and Algorithms, Advanced Programming (Python), Machine Learning, Deep Learning, Natural Language Processing, Software Engineering, Microcontrollers and Interface Technology, and Image Signal Processing.

English proficiency includes a CET-4 score of 453 and a pass at CLA-Basic level in the English Test for International Communication (ETIC).

Learning Experience

2024–2028 (expected)

Zero One Scholar Joint Training

Tsinghua University Qian Xuesen Program and Shenzhen X-Institute · ESRT, ORIC, and SURF training tracks

2026–Present

Research Visit at Shenzhen Bay Laboratory

Visiting student · Yaoqi Zhou’s group, Institute of Systems and Physical Biology · Downstream task optimization and multimodal extension of RESM

2025

Fudan Logic Summer School

Studied Foundations and Applications of Recursion Theory with Liang Yu. The course covered recursion theory, computable forcing, higher recursion theory, algorithmic information theory, and applications in areas such as descriptive set theory. Methods included priority arguments and effective transfinite induction.

Research & Engineering Projects

GPU-based streaming data acquisition for neutron spectrometers

Institute of High Energy Physics, Chinese Academy of Sciences · Supervised by Haiyun Teng · Individual project · 2025–2026

Built a GPU streaming acquisition prototype and its software framework, supporting DMA readout, AI image processing, data-stream output debugging, and runtime monitoring.

See the research experience at IHEP

MemoDroid · Dynamic hierarchical memory for LLM-based GUI testing

Institute of Software, Chinese Academy of Sciences · Supervised by Junjie Wang · Individual project · 2025–2026

Implemented a lightweight plug-in with episodic, reflective, and strategic memory and evaluated it on at least 200 real-world apps across 15 categories.

See the research experience at ISCAS

Plant counting from drone imagery

Undergraduate Innovation Training Program · Team lead · 2025–2026

Led work on a low-cost plant counting and replanting system. The design uses an adapted YOLOv8-tiny detector on low-altitude drone imagery and OpenCV to locate missing seedlings and map areas for replanting. It connects a geographic information system to a WeChat mini program.

The CV lists this project in its research phase. Planned outputs included one software copyright registration, one utility model patent application, and one article for a core journal.

Open-source processor design

Tsinghua Qian Xuesen Mechanics Class & Shenzhen X-Institute, X-Fusion · 2025 · Primary team lead

Recreated the RISC-I and Cray-I architectures with open-source electronic design automation (EDA) tools. We modeled the basic RISC-I logic in Logisim Evolution, implemented core modules in Verilog, prototyped RISC-V vector arithmetic with addition and subtraction instructions, and tested an open-source synthesis and place-and-route workflow.

The project received the X-Fusion Jingrui Award (零一菁锐奖).

Competitions & honors

·

China International College Students’ Innovation Competition

Core Team Member · Provincial Gold Award

In 2025, helped design a dashcam combining computer vision, deep learning, and edge computing. The system was designed to detect traffic lights, congestion, sudden lane changes, and hard braking, and send real-time text or voice alerts to following drivers.

·

Contemporary Undergraduate Mathematical Contest in Modeling (CUMCM)

Primary team lead · Provincial First Prize

In 2025, for male-fetus samples, the team modeled associations between gestational age, BMI, and the Y-chromosome fraction using mixed-effects models, decision tree regression, and Turnbull survival analysis, then applied TOPSIS to select test times across BMI groups. For female-fetus samples, the team used LASSO for feature selection and logistic regression for classification, with an entropy-weighted rule combining probabilities and Z-scores.

2025

7th Global Campus Artificial Intelligence Algorithm Elite Competition

AI + Integrated Circuits Track · Primary team lead · National Second Prize

Designed the Guangse Zhilian LiFi system using visible-light OFDMA, MIMO beamforming, and LED arrays to transmit data by modulating light. The design targeted smart-city and industrial IoT applications.

2025

13th National Collegiate Business Elite Challenge, Innovation and Entrepreneurship Competition

Retail Operations Simulation Track · Primary team lead · National First Prize

Analyzed Xinjiang Winka Times Department Store’s multiformat retail operations and local-market strengths, along with pressures from e-commerce, rising costs, and slow digitization. The team proposed ways to improve customer experience, optimize the supply chain, and introduce digital upgrades in stages.

2025

Mathematical Contest in Modeling (MCM)

Primary team lead · Successful Participant

Built a multi-objective model balancing tourism revenue, environmental pressure, and resident satisfaction to study overtourism in Juneau, Alaska. The team applied the model to Huangshan and proposed visitor management, economic diversification, and community engagement measures in a memo to Juneau’s tourism committee.

Outside research

I enjoy applying what I learn in class to practical projects and building prototypes from scratch. I captain the university table tennis and debate teams, and have received provincial recognition as a debater.

About Shenzhen X-Institute

Shenzhen X-Institute grew out of the Qian Xuesen Mechanics Class in Tsinghua University’s Xuetang Program. Founded in 2009, the class was the only one of 66 pilot programs in the Tsinghua Xuetang Program and the national program for top students in basic disciplines to focus on engineering fundamentals and interdisciplinary innovation across mechanics and engineering. Shenzhen X-Institute seeks to nurture innovators who use science and technology to benefit humanity, explore new ways of educating future innovators, and respond to the Qian Xuesen question of how to cultivate exceptional talent.