Projects
Projects Showcase
A selection of personal projects that demonstrate my technical skills and problem-solving ability in real-world scenarios.


Key features
- Product categorization by equipment type (extinguishers, smoke sensors, industrial systems, etc.)
- Smart product filtering and search
- Integrated shopping cart and checkout
- Quick contact and technical support form
Technologies
- WordPress, WooCommerce
Key features
- Fine-tuned language model using LoRA for resource-efficient training on small hardware (e.g., 4GB GPU)
- Processes natural language questions and provides accurate, localized responses in Vietnamese
- Modular CLI interface for interaction, easily extendable to REST APIs (FastAPI) or cloud deployment
- Supports containerization with Docker and orchestration via Kubernetes for scalable deployment
- Includes robust logging and error handling for Unicode and memory management across platform
Technologies
- Python 3.8+
- Hugging Face Transformers, PyTorch, PEFT (LoRA), Datasets, UTF-8 Logging
- Base model: DeepSeek-R1-Distill-Qwen-1.5B
Training Data
A JSON-based dataset containing real administrative and mental health Q&A examples, designed to reflect actual student inquiries.
Key features
- AI-powered mental health chatbot trained on real PDF, JSON and JSONL resources via Groq API
- Latest Grok 4 model for powerful prompts and coding.
- Commands for Khan Academy lookup and equation solving using SymPy
- YouTube music playback directly in Discord voice channels
- Scheduled and customizable content posting from Reddit, Pixiv, and VnExpress
Technologies
- Python 3.8+
- Discord API, Groq AI, Grok 4 API
- SymPy, Wikipedia API
- Reddit, Pixiv
Training Data
A JSONL-based dataset containing real administrative and mental health Q&A examples, designed to reflect actual student inquiries.