Curriculum Vitae

Undergraduate Researcher

AI Technology and Application Research Lab, FPT University

Conducting research in machine learning, deep learning, computer vision, and hand sign recognition to develop and enhance AI-driven solutions for real-world applications.

Bachelor of Science

FPT University, Ho Chi Minh City Campus

Artificial Intelligence

The Student Research Competition at FPT University

Third Prize

Awarded Second Prize for the research project titled 'Leveraging Modern Vision Architectures and Transfer Learning for Medical Image Classification'. The work explores the application of pre-trained deep learning models, such as ResNet, DenseNet, ViT, and ConvNeXt, in medical image classification using transfer learning techniques. By fine-tuning these models on small-scale datasets of CT, MRI, and histopathological images, the study achieved an average accuracy of 97.74% using a simple learning rate scheduling and early stopping strategy—demonstrating the effectiveness and generalizability of modern architectures in healthcare imaging tasks.

Led a research project applying transfer learning with state-of-the-art vision architectures (ResNet, DenseNet, ConvNeXt, ViT) for classifying medical images including CT scans, MRI, and histopathology. By fine-tuning pretrained ImageNet models and employing techniques like early stopping and learning rate scheduling, the project achieved an average classification accuracy of 97.74%—demonstrating the models' effectiveness in low-data medical imaging tasks.

Developing a gesture recognition system that extracts hand landmarks from video streams using MediaPipe and OpenCV, and models them using a hybrid architecture of CNNs (for visual features) and Graph Neural Networks (for topological gesture structure). The goal is to improve classification accuracy in dynamic hand gesture scenarios.

Built a full-stack web-based explainable AI (XAI) system integrating clinical text and medical imaging. Implemented model training, deployment pipelines, and user interface using Flask and Docker. The system allows users to input data and receive interpretable AI predictions through a streamlined web interface.