Hi, my name is Yiming Feng
I'm an AI/ML Engineer & Duke MSCS Student.

This project predicts the win probability in League of Legends matches using a deep neural network trained on in-game data (first 15 minutes). It also improves the model's local robustness through techniques such as PGD attacks, noise injection, Jacobian regularization, and TRADES.

A multimodal deep learning system for classifying 18 hand gestures using Graph Neural Networks, ResNet-34, and late fusion, achieving 99.72% accuracy on the HaGRID dataset.

Proactive AI Assistant is a Chrome extension that watches what you hover or select on the web and surfaces the most helpful AI tools in-place. It can explain code, summarize articles, graph equations, extract tables, run OCR on images, and more—all without leaving the page.

A deep learning project that classifies handwritten Chinese numerals using VGG16 and LeNet architectures, achieving ~97% accuracy on the Chinese MNIST dataset.

A real-time multiplayer Sichuan Mahjong web application with WebSocket-based gameplay, Google OAuth authentication, and Kubernetes deployment.

A full-stack stock trading simulation platform with real-time market data streaming, historical backtesting, and comprehensive portfolio management—designed to help users practice trading strategies without financial risk.