available for opportunities
Ziqi Gong
Machine Learning Engineer
focused on ML infrastructure and applied AI systems. I’m currently pursuing a Master’s in Electrical and Computer Engineering at the University of Michigan, specializing in data science and machine learning. I work on personal projects and research prototypes in CNNs and generative AI to deepen my understanding of real-world machine learning systems.
// education
Academic background
B.S. in Applied and Computational Mathematical Science
University of Washington
M.Eng. in Electrical and Computational Engineering
University of Michigan
// skills
What I work with
Languages
Java
Python
SQL
R
C#
Frameworks & Tools
Numpy
Pandas
PyTorch
scikit-learn
OpenCV
Git
// experience
Where I've worked
Research Assistant
July 2024 – Sept. 2024
HKUST - GZ
- Trained and optimized an object detection model using Roboflow and a compact, preprocessed dataset, achieving high detection accuracy while reducing data requirements.
- Developed a multimodal assistive system for blind users that converts visual information into audio and haptic feedback by integrating multiple AI agent APIs with custom hardware.
- Designed and conducted user experiments across multiple research projects, analyzing experimental data using statistical methods and data visualization.
- Collaborated with researchers to formulate research questions, develop interactive demos, analyze experimental results, and contribute to academic papers.
// projects
Things I've built
isThisFilm Image Classifier
Developed and deployed a ResNet-based CNN to classify film versus digital photography, improving accuracy from 60% to 86% through optimized preprocessing, augmentation, and architecture.
View on GitHub
K-in-a-Row Game-Playing AI
Developed a game-playing AI agent for K-in-a-Row using Minimax with Alpha-Beta pruning, heuristic evaluation, iterative deepening, and Zobrist hashing to improve strategic decision-making and search efficiency.
HanoCH
Developed a deployable accessibility system integrating computer vision, an LLM, and text-to-speech to support Blind users in tangible learnings.
View Images
LeGoDa
Developed an interactive AR learning system in Unity integrating visual, audio, and tactile interactions, and evaluated its impact through user studies, statistical analysis, and data visualization.
View Images