James Cui
Computer Science Student @ UW-Madison
Looking for Software Engineering Experience with a passion for machine learning, distributed systems, and building impactful technology. Currently exploring the intersections of AI and Software Engineering development.
About Me
I am a Computer Science student at the University of Wisconsin-Madison with a GPA of 3.93. My coursework in Advanced Algorithms, Software Engineering, and Deep Learning has provided me with a strong theoretical foundation, which I hope to use to resolve real-world problems.
I hope to learn something new everyday whether it be Machine Learning or Embedded Software. I've worked as a Peer Mentor and AI Research Assistant in hopes of one day being able to combine my experience in both fields to develop scalable and intelligent software.
Technical Skills
Languages
Frameworks & Libraries
Tools & Methodologies
Projects
Neural Machine Translator
Engineered a sequence-to-sequence translation system by fine-tuning a BART (Transformer) model, achieving a 4x improvement in BLEU score over baseline.
FCOS Object Detection
Implemented a Fully Convolutional One-Stage (FCOS) object detector from scratch, reaching 37.2% mAP on the PASCAL VOC validation dataset.
Smart Stock Inventory System
Led backend development for a full-stack retail/warehouse inventory system, featuring secure login and role-based access. Containerized the full application for consistent deployment.
Music Player Library
Led a team of 4 to create a custom data structure in Java to parse, sort, and visualize CSV data as a music playlist, ensuring quality with robust JUnit testing.
Experience
CS 506 (Software Engineering) Peer Mentor
University of Wisconsin-Madison
Jan 2025 - May 2025
- Co-led weekly tutorials for 200+ students on Agile, Maven, testing tools, and deployment workflows.
- Mentored 20+ students one-on-one on Git, Scrum processes, and Docker setup.
- Collaborated with TAs to create hands-on exercises, increasing student engagement in lab sessions.
Computer Vision Research Assistant
MadAbility Research Group
Jun 2024 - May 2025
- Annotated 1,000+ images and trained a YOLO model in PyTorch with 15% higher accuracy than baseline.
- Integrated OpenAI API to provide real-time personalized feedback from image analysis.
- Developed Unity C# AR barriers that improved OCD zone avoidance success rates by 60%.
- Researched object detection + AR strategies to support assistive technology for OCD patients.