Responsibilities include:

• Designed and conducted cognitive user studies with 24 participants using custom-built VR applications in Unity3D using the Meta Quest 3, collecting head/hand motion data to study subtle cognitive states.
• Built a reproducible research pipeline covering motion data collection (72 Hz), cross-platform time synchronization, video slicing, annotation alignment, and feature extraction, enabling machine learning analysis.
• Engineered and benchmarked classical ML and deep temporal models (SVM, Random Forest, LightGBM, LSTM, ResNet), demonstrating models can approach human-level performance in within-participant inference and highlighting cross-participant generalization as a key open challenge.
• First-authored papers presenting the first systematic evidence that VR motion telemetry encodes nuanced cognitive processes like confusion, hesitation, and readiness.

Responsibilities include:

• Proposed a self-training framework for 3D scene reconstruction that uses sparse input views and iteratively augments training with high-fidelity synthesized views, reducing NeRF’s reliance on dense view inputs.
• Led an independent research project focusing on time-series analysis, applying computer vision and deep learning techniques to extract and predict physiological signals from noisy laser-illuminated video data.

Responsibilities include:

• Developed a full medical imaging pipeline using U-Net segmentation and MMPose for skeletal keypoints estimation.
• Achieved a 75% (20° → 4.88°) reduction in measurement error through advanced image processing and model optimization, demonstrating expertise in computer vision and machine learning.
• Collaborated in a cross-institutional, interdisciplinary team with clinical experts to guide data annotation and ensure the real-world validity of the model’s outputs.
• Advisor: Prof. Xinhang Song

TCTM offers quality IT education targeting students aged three to eighteen through a hybrid platform combining live instruction, classroom tutoring, and online learning. The company was founded in 2002, formerly known as Tarena International, Inc.

Responsibilities include:

• Organized two intensive coding camps for 210+ students (ages 9–18) across 7 classes, delivering an engaging curriculum from foundational C++ to advanced algorithms and data structures, with real-world applications to strengthen students’ problem-solving skills.
• Enhanced communication and collaboration skills by regularly engaging with students’ parents to provide progress updates, and by working closely with fellow instructors to ensure high-quality, consistent instruction.