Origin.
I'm dedicated to leveraging fictional constructs to expand human capability—where anyone, regardless of geography or background, can experience practical skill-building without risk or cost through immersive simulations that augment real-world tasks from military training to medical drills.
Looking ahead, I envision creating a truly accessible Healthcare Metaverse—an AI-powered platform that brings high-fidelity medical training, diagnostics, and rehabilitation to anyone, anywhere. By combining real-time EEG analytics (for monitoring cognitive load or detecting early mTBI), camera-LiDAR 3D reconstructions (to build realistic virtual environments), and adaptive AI scenario engines, we could let trauma surgeons practice in dynamic, risk-free simulations; train first responders on disaster-response tactics; and even deliver personalized neuro-rehab exercises for patients at home. Running this on a clean-energy, AI-optimized infrastructure ensures it's both scalable and sustainable.
In short, I aim to use AI to bridge the gap between imagination and reality—democratizing critical healthcare skills and diagnostics through immersive, data-driven virtual worlds.
My Journey.
Research Focus
Advancing AI and machine learning research for biomedical applications and computer vision at UCI. My current work focuses on:
- Camera-LiDAR Fusion
Developing novel methods for depth map completion and 3D scene understanding
- EEG Biometrics
Enhancing identification accuracy with SOTA neural architectures for EEG-based person identification
- Transfer Euclidean Alignment
Improving transfer learning technique for mTBI detection across human and mouse EEG data
- Biomedical Signal Processing
Researching deep learning approaches for non-invasive fetal ECG monitoring
Key Milestones
Deep Learning for Fetal ECG Analysis
Research ContributionConducted comprehensive survey research on deep learning approaches for non-invasive fetal ECG monitoring, identifying key challenges and future directions for clinical applications.
Transfer Euclidean Alignment (TEA) for mTBI Detection
Research ContributionIntroduced a novel transfer learning technique that improved mTBI detection accuracy by 14.42% for intraspecies and 5.46% for interspecies datasets, enabling cross-species knowledge transfer from mouse to human EEG data.
Production Engineering Intern @ Meta x MLH
Industry ExperienceBuilt an open-source web application with Flask/NGINX/MySQL and implemented CI/CD pipelines and Prometheus/Grafana monitoring, reducing deployment time by 40% and incident response time by 50%.