A interactive demo showcasing real-time audio-reactive graphics, highlighting skills in game engine programming, shader development, and performance-conscious design.
Implemented a fully functional 32-bit RISC-V CPU in Logisim Evolution, including core modules such as the ALU (arithmetic, logic, shift, multiply), Register File with 32 registers (x0 hardwired to zero), Immediate Generator, Control Logic, and aligned Data Memory with write masks. The design was extended into a two-stage pipelined processor (IF, ID/EX/MEM/WB) with branch flush handling, supporting arithmetic, load/store, branch, and jump instructions.
A custom OpenGL rasterizer in C++ with point-in-triangle tests, bounding-box optimization, anti-aliasing (supersampling), and mipmap-based texture mapping for efficient, high-quality rendering.
A data science project using 200K+ Cook County housing records to build preprocessing pipelines and regression models, exploring feature engineering, predictive accuracy, and fairness issues such as multicollinearity and regressive taxation bias.
A machine learning project using the SpamAssassin dataset (8K+ emails) to build text preprocessing pipelines and logistic regression models, achieving ~89.5% accuracy while exploring precision/recall trade-offs, ROC analysis, and feature enhancements for improved spam detection.