Hi, I'm Will 👋

Full-stack 💻 & Machine Learning 🤖 & Security 🔐

Master student at Carnegie Mellon University 🎓. Passionate about new Techs 🚀✨.

Profile picture

About Me

I'm currently pursuing my Master of Science in Information Technology-Information Security at Carnegie Mellon University 🎓, with a strong background in software development and information security. My academic journey includes a Bachelor degree in EEE with Communication from the University of Glasgow 🎓 and a Bachelor degree in Communication Engineering from UESTC 🎓.

With experience in full-stack development, I have worked on various projects ranging from educational platforms to distributed systems. I am particularly interested in building secure, scalable applications and exploring the intersection of machine learning and software engineering.

When I'm not coding, you can find me exploring new technologies, contributing to open-source projects, or working on personal projects that combine my interests in machine learning and software development with security in mind.

Work Experience

Compani, Inc.

LinkedIn
Feb 2025 - Jun 2025

Co-Founder & Full-Stack Developer

  • Delivered MVP in 1 week, scaling to 500+ monthly active users with 0 major incidents; led the full product lifecycle from requirements to production.
  • Built cross-platform Flutter app with modules for application tracking, interview management, and user profiles, supporting real-time updates, offline caching, and AI response streaming.
  • Designed and deployed Spring Boot backend with PostgreSQL & Redis, integrating AWS Cognito & Google OAuth for secure, scalable multi-platform SSO.
  • Reduced LLM API costs 22% and cut timeout rate to <1% through caching and prompt templating.
FlutterSpring BootPostgreSQLRedisAWS CognitoGoogle OAuthLLM API

GoodRunss AI, Inc.

LinkedIn
Jul 2025 - Aug 2025

Frontend Lead

  • Architected the frontend of an AI-powered court management dashboard, implementing modular UI components with shadcn/ui and state management for scalable feature growth.
  • Built responsive, high-performance Next.js UI with real-time court availability tracking, interactive Chart.js visualizations, and optimized rendering via SSR, code-splitting, and lazy-loading.
  • Managed collaborative Git workflows (feature branching, PR reviews) and containerized deployments via Docker on AWS EC2, cutting deployment time under 5 mins and achieving 99.9% uptime.
  • Integrated REST APIs with backend AI models to deliver busy-level predictions, enabling data-driven scheduling decisions and boosting user engagement by 15%.
Next.jsGitDockerAWS EC2

My Projects

Kitefun

Educational Platform

Co-founded Kitefun, developing an educational platform with Django, Flutter, and AWS infrastructure.

DjangoFlutterAWSPostgreSQL
5G Base Station

5G Configuration Platform

Built a secure control dashboard for 5G base station configuration using Next.js and FastAPI.

Next.jsFastAPIWebSocketDocker
Raft Consensus

Distributed System

Implemented Raft in Go with leader election, log replication, and fault-tolerant state machine replication; built custom RPCs using goroutines and channels to simulate failures and validate correctness under concurrency. Tested log consistency, leader stability, and fault resilience in a simulated cluster, gaining hands-on experience with distributed systems concepts such as consensus and the CAP theorem.

GoDistributed SystemsRaft
Activity Recognition

Machine Learning Pipeline

Developed a supervised ML pipeline for wrist-worn accelerometer data with segmentation, windowing, normalization, and feature extraction. Trained and compared Random Forest, SVM, and CNN, achieving 90%+ test accuracy and deploying the best model via Flask API for mobile/wearable integration.

PyTorchscikit-learnFlaskRandom Forest
Smart Surveillance

Object Detection System

Built a real-time surveillance system using YOLOv10 for pedestrian and vehicle detection with 92% mAP.

YOLOv10FastAPIReact
Smart Vehicle

Multi-functional Vehicle

Led a 10-member team in designing a smart vehicle capable of line detection, radar-based obstacle sensing, PID-controlled movement, and automated ball throwing. Built OpenMV visual recognition module and STM32, achieving 98% detection accuracy and ranking 1st/100+.

PythonCOpenMVSTM32PID

Skills & Technologies

P
Python
J
JavaScript
T
TypeScript
G
Go
J
Java
C
C/C++
D
Django
R
React
N
Next.js
F
Flutter
S
Spring Boot
P
PyTorch
T
TensorFlow
A
AWS
D
Docker
P
PostgreSQL
M
MongoDB
R
Redis
G
Git
C
CI/CD
W
WebSocket
O
OAuth
J
JWT
S
STM32