CV

Teaching Assistant

AI VIETNAM

Developing curriculum and teaching materials for the AIO program, covering fundamental Machine Learning, Deep Learning, and advanced topics such as AI Agents and LLM Reasoning.

• Created comprehensive educational content for AI/ML courses
• Designed exam questions for Linear Algebra, Probability, Statistics, ML, and DL
• Teaching materials available at: https://drive.google.com/drive/folders/1JTLMNDsnOKD4Yo1e2x3CZVMUXkbP8KD2

B.Sc. in Computer Science

VNU-HCM, University of Science

Computer Science (High-Quality Program)

• Multi-modal AI
• Deep Learning
• Machine Learning
• Data Science Programming

Dean's List Honoree (5 Semesters)

VNU-HCM, University of Science

Recognition for academic excellence with GPA 3.98/4.0

An Automated Pipeline for Constructing a Vietnamese VQA-NLE Dataset

Proceedings of the Fifth International Conference on Intelligent Systems and Networks (Lecture Notes in Networks and Systems, Springer Nature Singapore)

Officially published ViVQA-X dataset paper detailing an automated, multi-LLM pipeline that constructs the first Vietnamese Visual Question Answering dataset with natural language explanations.

Describe Anything Model for Visual Question Answering on Text-rich Images

ICCV Workshop 2025

Multi-modal VQA approach demonstrating strong performance across six benchmark datasets through comprehensive evaluation of Vision-Language Models.

Counterfactual Reasoning for Robust Visual Question Answering

Graduation thesis proposing counterfactual training strategies, batch-contrastive losses, and curriculum learning to mitigate language bias in VQA.

• Three-stage curriculum for stable optimization
• Batch-contrastive loss with counterfactual samples
• State-of-the-art performance on VQA-CP v2 (annotation-free methods)

Heineken Image Analysis Tool

Backend and model engineering for a multi-model AI pipeline that automates brand compliance and safety analysis from images.

• FastAPI inference service
• Integrated YOLOv10, Owlv2, PaddleOCR, and CLIP
• Delivered actionable reports for on-site audits

AI-Powered Math Solving Assistant

Multi-modal chatbot that solves math problems through Gemini API fine-tuned on MetaMathQA-40K with OCR support for handwritten inputs.

• Vertex AI fine-tuning workflow
• OCR integration for image inputs
• Web interface for interactive tutoring

Manga Popularity Prediction

End-to-end data science pipeline for predicting manga scores using scraped metadata, exploratory analysis, and ensemble regression.

• Automated web scraping and preprocessing
• Compared Linear Regression, Random Forest, and XGBoost
• Bayesian Optimization for hyperparameters