I'm an AI/ML engineer and NLP researcher specializing in automatic speech recognition (ASR), accent adaptation, and noise-robust deep learning systems. I recently graduated with a Bachelor of Computer Science (Artificial Intelligence) from Swinburne University of Technology (GPA 3.8/4.0 US scale, High Distinction) and hold an Associate of Science in Engineering from Highline College, Washington, USA.

I am the lead researcher and first author on three published/submitted papers in ASR, with all experimental design, model training, and evaluation done independently. My work has been recognized with a Best Paper Award at ICISN 2026 (Springer).

Journal paper — under submission

• Audio-visual noise detector: 1.67M params, 99.4% accuracy, ROC-AUC 0.9995, 2.65ms inference
• Accent classifier: 99.9% accuracy
• LoRA-adapted Whisper-Small achieves 10.08% WER, outperforming Whisper-Medium (12.32%) using only 0.36% of its parameters
• Validated on real YouTube videos | All results: p < 10⁻¹³

Published by Springer at ICISN 2026 (Hanoi, Vietnam) — Best Paper Award

• 17% relative WER reduction: 14.98% → 12.43%
• CNN accent detector: 99.82% accuracy in < 50ms, no fine-tuning required
• t = 11.32, p < 0.001, Cohen's d = 0.21

Published by Springer at CITA 2025 — DOI: 10.1007/978-3-032-00972-2_31

• WER reduced: 18.08% → 4.75%
• Scottish English WER: 18.18% → 9.09% via phonetic normalization
• BLEU score: 0.83 → 0.91

Graduation Capstone · Team of 7 · My role: AI Engineer

An intelligent rehabilitation platform for elderly users, delivered to a real client across 8 Agile sprints.

• AI modules I built: TTS voice guidance (pyttsx3), pain recognition (MediaPipe Face Mesh, 6-state, 85% CPU reduction), AI chatbot with contextual response generation
• 91.7% pose estimation accuracy using MediaPipe
• Stack: FastAPI · React · TypeScript · MediaPipe · WebSocket · MySQL · Docker · JWT

Specializations: ASR · NLP · LoRA / PEFT · CNN · ResNet · Ensemble Methods · Generative AI · Prompt Engineering

🇻🇳 Vietnamese (Native) · 🇺🇸 English (Fluent — IELTS 7.5–8.0 equivalent) · 🇰🇷 Korean (Basic — TOPIK I)