Electronic and computer engineering graduate with strong AI skills and several development roles, able to read product and application needs from both the hardware and the software side.
Studied abroad, with experience communicating and collaborating across cultures.
During my internship I worked with local operations teams in Europe to develop and ship B2B marketing solutions; as a teaching assistant and in product demos I gained experience explaining technology, answering questions and presenting. I want to combine engineering analysis with an understanding of the business.
MS EE/CE Honors Students 2026 ↗To recognize MS EE students with an exceptional academic record, active EE MS students with a GPA in the top 10% of the class are included in the MS Honors program.
Courses (8)
IOT - INTELLIG & CONNECTED SYS
FORMAL VERIF HW SW SYSTEM
COMPUTER NETWORKS
COMPUTER VISION II: LEARN
DATA CENTER PROCESSING
MODELING & PERFORMANCE EVALUATION
LLM BASED GENERATIVE AI
FUND-LARGE-SCALE DISTRIBUTED SYSTEMS
Zhejiang University
Zhejiang University–University of Illinois Urbana-Champaign Institute
B.Eng. in Electronic and Computer Engineering | ZJU–UIUC dual-degree program | GPA: 3.8 / 4.0
ThesisImage reconstruction with generative models
TeachingTeaching assistant for Discrete Mathematics: organised teaching activities, prepared course materials, and followed up on, collected and answered student questions
Courses (23)
Introduction to Data Mining
Senior Thesis
Introduction to Robotics
Senior Design Project Lab
Digital Systems Laboratory
Digital Signal Processing
Introduction to Algorithms & Models of Computation
Machine Learning
Applied Parallel Programming
Artificial Intelligence
Communication Networks
Computer Systems Engineering
Game Development
Probability with Engrg Applic
Database Systems
Analog Signal Processing
Data Structures
Linear Algebra with Computational Applications
Basic Discrete Mathematics
Computer Systems & Programming
Introduction to Computing
Intro to Computing: Engrg&Sci
Introduction to Electronics
Experience
Alibaba International Digital Commerce Group AI Application Development Engineer
Merchant acquisition (B2B marketing)
Cross-region collaboration: Worked with local operations in Europe, dug into e-commerce merchants' business scenarios, surfaced pain points and produced technical solutions
Merchant efficiency and traffic: Owned end-to-end development of a smart marketing assistant for merchants, designing and shipping automated product selection, pricing and submission that supports automation for tens of millions of SKUs a day; built an activity-page routing strategy from merchant profiles to recommend suitable campaigns, reaching a 51.5% click-through conversion among target merchants and meeting the business goal
Solution playbook (Agent Loop showcase): Built a one-stop AI coding showcase on the internal platform; led development of the integration-test, campaign-creation and registration Skills, connecting an end-to-end automated testing flow and closing the development loop, with an overall efficiency gain of about 30%
Alibaba International hackathon: Built a conversational agent in 48 hours for an emotion-driven e-commerce entry point; the project reached the second round (100/1388), and I designed materials, presented on site and gave the product pitch, winning a Top 25 popularity award
Hikvision Digital Technology Embedded Software Engineer
On-site delivery and quality control: Collected product data on the factory floor and identified needs for technical improvements
Embedded development: Developed and integrated core modules of a commercial cleaning robot's main control system on Linux, and sensor control and data acquisition logic on FreeRTOS, keeping the microcontroller's control precise and real-time
In a team of four, designed and built an end-to-end simulation workflow driven by real inference and training traces, enabling reusable energy-aware performance analysis of large deep-learning systems.
Implemented collective-communication and kernel-level event capture and format conversion for multi-GPU distributed fine-tuning (FSDP + LoRA), covering several mainstream models.
Built a web application for US equity decisions that addresses serial latency in multi-agent systems, decision transparency and ease of use.
Used a Gin gateway with a FastAPI inference service, persisting reasoning trails and decisions in PostgreSQL (GORM); containerised the core microservices with Docker for fast full-stack deployment.
Implemented parallel multi-agent execution and dependency control with LangGraph, cutting end-to-end decision latency in multi-agent scenarios by about 70% compared with a serial pipeline, with a choice of multiple APIs, local Ollama or OpenClaw for inference.
Developed a Plants vs. Zombies game on an FPGA, integrating hardware modules, a Nios II soft-core processor and peripherals; ran ModelSim simulation and system verification for hardware–software co-design of display, input and control logic
Optimised LeNet-5's convolution layers in CUDA with streams, Tensor Cores, shared/constant-memory tiling, quantisation and loop unrolling to raise throughput: 27,909× faster than a naive CPU implementation and 36% faster than an unoptimised parallel baseline.
In a team of four, built a Linux-like OS kernel in C and x86 assembly.
Developed virtual memory, a file system, terminal display, interrupts, system calls, exception handling and device drivers (keyboard, RTC, PIT), delivering multi-process switching, multi-terminal management and scheduling with kernel and user space isolated.
Framed PET (positron emission tomography) reconstruction as a Bayesian inference problem and implemented a deep probabilistic imaging pipeline, improving reconstruction accuracy and providing the uncertainty estimates that clinical diagnosis relies on.
Applied normalizing flows (RealNVP, Glow) to model the posterior distribution of tracer activity and scaled training with multi-GPU parallelism.
Developed a generative-AI plugin for Blender that automates the 3D design workflow: prototype management, segmentation and Gaussian–mesh conversion.
Built a text-to-3D generation system on Transformer/Diffusion models combined with Gaussian splatting and mesh rendering to simplify modelling; deployed a Flask backend that records user interactions and serves real-time 3D generation.
Bayesian PET Reconstruction with Learned Normalizing Flow Priors
IEEE International Symposium on Biomedical Imaging (ISBI) 2026Second author
For medical image reconstruction, modelled the prior and posterior distributions of tracer activity with Flow Matching / Normalizing Flows, accelerated training across multiple GPUs on a Slurm cluster, and achieved reliable uncertainty quantification and high-quality reconstruction
Learned the basics of financial markets and asset allocation through finance courses; took part in an asset-allocation project, designing and analysing an allocation plan and presenting it, strengthening my applied financial knowledge, teamwork and ability to present a plan.
