Projects

Selected Work

Technical and research projects across NLP, AI Safety, computer vision, explainability, and political analysis. I care most about work that helps me connect abstract ideas to real systems, and real systems back to deeper questions.

2025·Ongoing

Rubric-based Grading Tool

A benchmark-oriented grading workflow for reasoning-heavy CS302 exam questions using Gemini, prompt experiments, score-alignment metrics, semantic similarity, and a Streamlit dashboard.

2025·Project

CoherentEyes Fake News Detection

A public-facing project on misinformation, information integrity, and generative model vulnerabilities, with SHAP-based interpretability analysis across classical and deep learning models.

2025·Project

Truyện Kiều Project

A Vietnamese poetry generation and authorship attribution project using Transformer, GPT-2, TF-IDF, and SVM.

2025·Project

Stock Price Prediction

A time-series forecasting project using TensorFlow/Keras, market indicators, and portfolio-oriented analysis.

Working style

I'm drawn to projects that let me move between technical detail and broader stakes. Sometimes that means building systems, sometimes it means formalizing a political problem, and sometimes it means treating a project as a way to learn how to ask better questions.