Mechanical Engineering student @ UofT with a minor in Artificial Intelligence Engineering. On the computer-science side of things, I'm currently working on AI/ML systems focused on Vietnamese NLP, multi-modal AI, and real-time computer vision. Currently leading the opto-mechanical subsystem at UTAT Space Systems and exploring how AI transforms the more traditional mechanical engineering workflows and tasks—feel free to reach out.
|
End-to-end vision-language pipeline: BLIP generates English captions → CLIP ranks by image-text similarity → NLLB translates to Vietnamese. Evaluated with BLEU/METEOR.
|
LSTM-Autoencoder detecting anomalies in NASA SMAP/MSL satellite telemetry (temperature, voltage, gyroscope). 92%+ F1-score with reconstruction error thresholding.
|
|
Fine-tuned NLLB-200 for trilingual Vietnamese ↔ English ↔ Japanese translation with beam search decoding and language auto-detection. Gradio interface.
|
Fine-tuned OpenAI Whisper on Mozilla Common Voice Vietnamese. Audio preprocessing with Librosa (resampling, noise augmentation, silence trimming). WER/CER evaluation.
|
|
Fine-tuned VinAI/PhoBERT for 3-class sentiment classification on 10,000+ Vietnamese product reviews. 85% accuracy. Deployed on HuggingFace Hub.
|
Real-time hand tracking with MediaPipe mapping 21-point landmarks to 10 custom gestures for desktop control (cursor, volume, scroll). 30+ FPS on CPU.
|
Open to ML/AI internship opportunities — especially in NLP, computer vision, and multi-modal systems.