3rd-year CS Student @ HCMUS · Data Science & AI Automation · Building AI-ready data pipelines and end-to-end ML systems

I'm a 3rd-year Computer Science student at Ho Chi Minh University of Science (HCMUS), focused on building practical AI and data systems that connect raw data, analytics, machine learning, and automation.

I enjoy working on projects that go beyond notebooks: collecting and cleaning data, designing feature pipelines, building APIs, creating dashboards, containerizing applications, and turning models into usable products.

• 🔭 Recent focus: AI-ready data pipelines, analytics agents, realtime monitoring systems, and ML applications
• 🌱 Currently learning: MLOps, CI/CD, data engineering fundamentals, LLM tool-calling, and scalable model serving
• 💡 Interested in: Applied ML, AI Automation, Data Science, NLP, Explainable AI, and production-oriented AI systems
• 🎯 Next goal: Land a Data Science / AI / Automation Internship in 2026

An AI-assisted analytics platform that allows users to upload tabular datasets, profile data quality, generate KPI dashboards, and ask grounded natural-language questions over data.

• Built CSV/XLS/XLSX upload, dataset profiling, missing value checks, duplicate detection, column type inference, and metadata storage.
• Implemented semantic column mapping for business roles such as revenue, date, category, quantity, profit, campaign, and department.
• Developed deterministic Pandas-based analysis tools with an AI Copilot layer for reliable data question answering.
• Designed a production-style structure with FastAPI backend, React dashboard, Docker Compose, tests, scripts, and documentation.

FastAPI Pandas React SQLite Docker AI Copilot Data Quality Analytics Dashboard

A near real-time environmental intelligence system for monitoring AQI and weather data across Vietnam, combining data pipelines, analytics, forecasting, anomaly detection, and dashboard reporting.

• Designed an end-to-end data flow from data collection, streaming, storage, analytics, forecasting, insight generation, and reporting.
• Built dashboard views for AQI monitoring, maps, alerts, forecast trends, province comparison, and environmental insights.
• Integrated a realtime-oriented architecture using API services, streaming concepts, time-series storage, and frontend visualization.
• Focused on moving beyond passive monitoring by adding short-term prediction, anomaly detection, and explainable insights.

Python FastAPI React Kafka TimescaleDB AQI Forecasting Anomaly Detection

A fashion e-commerce forecasting project focused on predicting daily Revenue and COGS using multi-source business data.

• Processed multi-source datasets including orders, payments, products, customers, promotions, inventory, web traffic, shipping, returns, reviews, and geography.
• Built feature engineering workflows for time-series forecasting while avoiding data leakage.
• Conducted revenue trend analysis, seasonality analysis, promotion impact analysis, inventory risk analysis, and RFM customer segmentation.
• Compared multiple forecasting approaches including statistical baselines, LightGBM, XGBoost, CatBoost, Prophet, and explainability techniques.

Pandas Time Series Feature Engineering LightGBM XGBoost CatBoost Prophet SHAP

An end-to-end cardiovascular risk assessment system with ML inference, explainability, OCR, LLM-assisted report interpretation, and containerized deployment.

• Built a machine learning pipeline using preprocessing, feature engineering, XGBoost training, cross-validation, and model evaluation.
• Integrated SHAP explainability to make predictions more interpretable.
• Developed a FastAPI backend and Streamlit frontend for interactive prediction and report generation.
• Added Docker-based deployment, API validation, logging, and GitHub Actions CI/CD workflow.

XGBoost SHAP FastAPI Streamlit Docker GitHub Actions Gemini API EasyOCR

An automated NLP pipeline for parsing LaTeX scientific papers and matching in-text citations with bibliography references.

• Built a one-command pipeline for paper parsing, reference extraction, candidate generation, feature engineering, model training, prediction, and evaluation.
• Combined TF-IDF, Jaccard similarity, Sentence-BERT embeddings, and Random Forest classification.
• Engineered 44 features and evaluated ranking performance using MRR and Recall@K.
• Improved citation matching performance over a TF-IDF heuristic baseline.

Python NLP Sentence-BERT TF-IDF Random Forest LaTeX Parsing Automation Pipeline

A Vietnamese hotel recommendation system that matches user preferences with hotel features using content-based filtering, semantic search, and location-aware matching.

• Built a recommendation workflow using hotel attributes, facilities, location information, text features, and user preferences.
• Implemented semantic search and Named Entity Recognition to understand Vietnamese search queries.
• Processed and transformed hotel data for downstream recommendation tasks.
• Designed a user-friendly search experience for Vietnamese travelers.

Recommendation System Semantic Search NER Content-Based Filtering Pandas Scikit-learn

• ⚙️ MLOps: experiment tracking, model registry, CI/CD, containerized model serving
• 🧠 LLM Engineering: tool calling, structured output, RAG, evaluation, and hallucination control
• 🏗️ Data Engineering Basics: data pipelines, validation, orchestration, warehouse/lakehouse concepts
• 🚀 Production AI Systems: FastAPI, Docker, monitoring, scalable inference, and automation workflows

• 💼 Data Science / AI / Automation internship opportunities
• 🤝 Collaborating on ML, NLP, analytics, and AI automation projects
• 💬 Discussing ideas around applied ML, data pipelines, explainable AI, and production-ready AI systems

Feel free to reach out via LinkedIn or email!