@@ Third-year AI Student @ FPT University | GPA: 8.62/10.0 @@
+ Published Researcher in Springer LNAI (Q2)
+ Founder of SpeedyLabX Research Group
! Seeking Research Internships & Collaboration Opportunitiesfocus: "Building transparent, interpretable AI systems"
specialization:
- Explainable AI (SHAP, Attention Mechanisms)
- Multimodal Learning (Audio-Text-Vision)
- Time-Series Forecasting (Transformers)
- Speech Processing (Low-Resource ASR)I'm an AI researcher passionate about making machine learning models more interpretable and trustworthy. As founder of SpeedyLabX, I lead a team of 10 undergraduate researchers exploring cutting-edge problems in explainability, multimodal fusion, and speech systems.
Air Quality Forecasting metrics = {
"MAE": 0.7470,
"R²": 0.9545,
"improvement": "35.9% vs XGBoost"
}Transformer-based system with SHAP explainability for public health warnings
AJCAI 2025 • First Author |
Emotion Recognition architecture = {
"audio": "Wav2Vec2",
"text": "RoBERTa",
"fusion": "GATv2"
}Graph-based multimodal fusion with intrinsic attention explainability
IEMOCAP • RAVDESS • MELD |
Speech Recognition model = {
"params": "5.4M",
"data": "745 hours",
"WER": "33.09%",
"CER": "15.28%"
}Parameter-efficient CNN-BiLSTM with CTC for Vietnamese speech
Competitive vs Foundation Models |
@inproceedings{le2025smokenet,
title = {Proactive Air Quality Forecasting and Health Alert System for Melbourne},
author = {Gia-Hung Nguyen Le and others},
booktitle = {AI 2025: Advances in Artificial Intelligence},
series = {Lecture Notes in Artificial Intelligence (LNAI)},
publisher = {Springer Nature Singapore},
year = {2025},
note = {Q2 (SJR 2024)}
}📄 Full Paper • First Author & Corresponding Author
research_focus:
time_series:
- Transformer architectures
- Environmental prediction
- Chronological cross-validation
multimodal_learning:
- Audio-text fusion
- Graph Neural Networks
- Conversational AI
explainable_ai:
- SHAP interpretability
- Attention mechanisms
- Integrated Gradients
speech_processing:
- End-to-end ASR
- Low-resource languages
- CTC optimization |
technical_stack:
languages:
- Python
- SQL (PostgreSQL)
frameworks:
deep_learning: [PyTorch, TensorFlow, Keras]
ml_tools: [Scikit-learn, Hugging Face, OpenCV]
mlops:
- Docker
- Weights & Biases
- Git
- Linux
- Streamlit
data_science:
- Pandas
- NumPy
- Matplotlib |
|
Languages
|
Deep Learning
|
ML Ecosystem
|
MLOps & Tools
|
| 🎯 Achievement | 📅 Year | 🔗 Details |
|---|---|---|
| Published in Springer LNAI | 2025 | First-author paper at AJCAI (Q2) |
| Founded SpeedyLabX | 2024 | Research group of 10+ undergraduates |
| Certificate of Merit | 2024-25 | Honorable Student (Semesters 3, 4, 5) |
| Vietnamese ASR Breakthrough | 2025 | 33.09% WER on 745-hour corpus |
| IBM Data Science Certifications | 2024 | Professional Certificate track |
const speedyLabX = {
role: "Founder & Student Lead",
established: 2025,
team_size: "10+ undergraduate researchers",
activities: [
"Weekly technical workshops and paper reading groups",
"End-to-end publication workflow management",
"Mentoring on deep learning & research methodology",
"Successfully guided first AJCAI-accepted paper"
],
focus_areas: [
"Explainable AI",
"Multimodal Learning",
"Speech & NLP",
"Time-Series Forecasting"
]
}! Actively seeking Research Internships & Collaborations in:
+ Explainable AI & Interpretable Machine Learning
+ Multimodal Learning (Audio-Visual-Text)
+ Speech & Natural Language Processing
+ Time-Series Forecasting & Graph Neural Networks
Academic Profiles
FPT University • Ho Chi Minh City, Vietnam 🇻🇳