Xuan Truong Nguyen
Data Science | Data Engineering Enthusiast
About Me
I'm a Data Science enthusiast with a passion for turning complex problems into elegant solutions. My approach? Finding the unconventional path that others might overlook.
Technical Skills
Programming Languages
Machine Learning & Data Science
Data Engineering & Cloud
Experience
Data Engineering Intern
Lions Eye Institute — Jun 2025 – Present
- Developed an end-to-end data transformation pipeline to standardize over 250,000 eye health records collected over 30 years (~60K annual visitors) using an international healthcare data model
- Cleaned, integrated, and mapped clinical data for accurate analysis and research compatibility across multiple systems
- Followed established conventions and vocabularies to support global research standards and interoperability
- Collaborated with clinical and technical teams to validate mappings and ensure data quality
Geospatial Data Science Intern
The Wilderness Society — Feb 2025 – Jun 2025
- Automated annual land cover classification (1987–2025) using unsupervised K-Means clustering on Landsat imagery with Remotior Sensus
- Conducted remote sensing analysis with Google Earth Engine and Sentinel-2/Landsat to detect long-term deforestation trends
- Generated time-series land cover maps and vectorized outputs to quantify and visualize mining footprint expansion across decades
- Analyzed spectral changes and built geospatial datasets to assess environmental impact across decades
Projects
Cross-Modal Transformer with Adversarial Domain Adaptation for Robust Emotion Recognition
Personal Project — Mar 2025 – Aug 2025
- Developed a state-of-the-art deep learning model for multi-class emotion recognition (5 emotions) from concurrent EEG (Differential Entropy features) and eye-tracking data, utilizing the SEED-V dataset
- Designed and implemented a custom Transformer-based architecture featuring dedicated EEG and eye-tracking encoders, Cross-Modal Attention mechanism, and Feature Importance Modules
- Integrated a robust adversarial domain adaptation strategy using a Gradient Reversal Layer (GRL) and subject-specific normalization layers to mitigate inter-subject variability
- Validated model generalization using rigorous Leave-One-Subject-Out (LOSO) cross-validation across 16 subjects, achieving a mean accuracy of 75.42% ± 11.15% and a best-fold accuracy of 91.11%
Predictive Weather Analysis Using Web Scraping, ML Algorithms, and Time Series Modeling
Personal Project — Jul 2024
- Engineered a web scraping tool using Selenium, successfully collecting 12 years of historical weather data for comprehensive analysis
- Achieved 99% accuracy in weather forecasting by implementing a prediction system using KNN, Random Forest, and XGBoost algorithms, leveraging features like humidity, dew point, and pressure
- Conducted time series analysis using SARIMAX model to predict future temperature trends, enhancing the accuracy of short-term weather forecasts
Contact
Get In Touch
I'm always open to discussing new opportunities, collaborations, or just having a chat about data science and technology.