cv

Research Student

BKAI Laboratory (Bioinformatics Research Team)

Working with the Bioinformatics team on the Drug Response Prediction project. Developed novel multi-level information integrating fragment-level and atom-level GNN paradigms.

• Proposed a novel paradigm bridging atom- and fragment-level representations using compact chemical vocabulary.
• Achieved state-of-the-art performance on standard CCLE dataset for within-dataset drug response prediction.

Research Student

International Research Institute MICA

Conducting research on Traumatic Brain Injury (TBI) Classification using multimodal deep learning and NLP techniques.

• First author of 1 journal paper and 1 conference paper.
• Developed viTBI-BERT, a Vietnamese Language Model for TBI prediction.

Master

Institut Polytechnique de Paris

Data Analysis and Pattern Recognition

Bachelor

Hanoi University of Science and Technology

Biomedical Engineering

• Introduction to Biology (A+)
• Medical Image Processing (A)
• Bioinformatics (A+)

High School

Nguyen Hue High School for the Gifted

Physics

Best Presentation Award

Electronics and Electrical Engineering Council, HUST

Recognized for delivering the top-ranked thesis presentation among all candidates in the Electronics and Electrical Engineering department.

Second Prize - Health Systems Innovation Hackathon

Hanoi Medical University, Vietnam Hub

Building high-value health systems through Artificial Intelligence.

Third Prize - HUST Annual Student Research Conference

Hanoi University of Science and Technology

Awarded for research excellence in the track of AI Applications, Blockchain, and Big Data.

Scholarships for Outstanding Academic Performance

Nguyen Hue High School for the Gifted

Received scholarships in 2019, 2020, and 2021.

TBI-TTM: Traumatic Brain Injury Prognosis with Textual Data under Missing Tabular Conditions

The 24th International Symposium on Communications and Information Technologies (ISCIT 2025)

First author. Proposed a method to enhance robustness with missing data without imputation, increasing sensitivity by approximately 10%.

viTBI-BERT: A Vietnamese Language Model for Prediction of Traumatic Brain Injury

Journal on Information Technologies & Communications, Vol. 2025, No. 1, pp. 54-67

First author. Developed a domain-adapted BERT model for TBI prediction using Vietnamese physician notes.

Multimodal Traumatic Brain Injury Prognosis Assessment

Developed an end-to-end multimodal deep learning pipeline to predict TBI severity by fusing structured clinical data with unstructured physician notes from 503 patients. The model outperformed uni-modal baselines by 1–2% across four key metrics.

• Fine-tuned a LLM domain-adapted BERT model achieving 71% sensitivity.
• Used adjusted Transformer encoder to handle missing data without imputation.

Drug Response Prediction System

Proposed a novel multi-level information integrating fragment-level and atom-level GNN paradigm. Bridges atom- and fragment-level representations using a compact chemical vocabulary and principal subgraph mining.

• Achieved state-of-the-art performance on standard CCLE dataset.
• Focused on within-dataset drug response prediction.

Genomic Variant Analysis and Annotation

Analyzed germline and somatic variants using GATK. Practiced workflows for variant preprocessing, discovery, refinement, and evaluation.

• Utilized tools: GATK, SAMtools, Picard, IGV.
• Performed analysis using R and Python in Linux environment.