Working on Gaussian Splatting at Meta Reality Labs.

Research in computational imaging, displays, and optics.• Large Étendue 3D Holographic Display with Content-adpative Dynamic Fourier Modulation: SIGGRAPH Asia 2024• Full-colour 3D Holographic Augmented-Reality Displays with Metasurface Waveguides: Nature 2024• Holographic Parallax: SIGGRAPH Emerging Technologies 2024• High-brightness Holographic Projection: Optics Letters 2023• Neural Holographic Displays for Virtual Reality: SIGGRAPH Emerging Technologies 2023

• Performed benchmarking, visualization, and analysis on neural PDE solvers.• Designed an algorithm to optimize light field displays with deep learning.• Worked on cardiovascular disease prediction using large-scale clinical data. Improved the prediction AUC score (0.81 ROC-AUC) by over 40% compared to the baseline knowledge of doctors.

Research in light fields, computational imaging and displays, view synthesis, neural radiance fields- Light Field Synthesis and Computational Light Field Displays⠀• Proposed the coded time-division multiplexing display concept that improves light field display quality by 14% in terms of peak signal to noise ratio (PSNR), compared to baseline integral imaging display designs.⠀• Proposed an algorithm to synthesize a light field from a pair of stereo images.⠀• Research resulted… Show more Research in light fields, computational imaging and displays, view synthesis, neural radiance fields- Light Field Synthesis and Computational Light Field Displays⠀• Proposed the coded time-division multiplexing display concept that improves light field display quality by 14% in terms of peak signal to noise ratio (PSNR), compared to baseline integral imaging display designs.⠀• Proposed an algorithm to synthesize a light field from a pair of stereo images.⠀• Research resulted in a first-authored paper accepted to IEEE ICIP 2021, a first-authored paper accepted to IEEE TIP, and a first-authored paper submitted to IEEE TCSVT- Student Reviewer⠀• Served as the reviewer for the following conferences and journals: IEEE ICIP, TIP, ICASSP, ICME Show less

Worked on open source projects with Google Mentors and other SPS participants in the summer of 2020 (2020 Google SWE canceled due to COVID-19).• Developed a full-stack image processing web application from scratch using Flask, pure JavaScript/jQuery/HTML/CSS, and Google Cloud. • Leveraged numpy and scipy to support fast image processing algorithms implementations such as Kmeans++, Median Cut, and Otzu's Method for image quantization, bilateral filters, and etc .

• TA for Machine Learning, Spring 2020 (Instructor: Prof. Hung-Yi Lee)• TA for Signals and Systems, Spring 2020 (Instructor: Prof. Lin-shan Lee)

Unsupervised representation learning, self-supervised learning, and generative adversarial networks (GAN) applications- Unsupervised Representation Learning / Contrastive Learning⠀• Designed a fully-differentiable augmentation policy searching scheme to search for optimal augmentation policies that benefit contrastive learning on 2D images.⠀• Worked on contrastive learning on 3D representations to learn robust features for various kinds of 3D data (voxels, point clouds… Show more Unsupervised representation learning, self-supervised learning, and generative adversarial networks (GAN) applications- Unsupervised Representation Learning / Contrastive Learning⠀• Designed a fully-differentiable augmentation policy searching scheme to search for optimal augmentation policies that benefit contrastive learning on 2D images.⠀• Worked on contrastive learning on 3D representations to learn robust features for various kinds of 3D data (voxels, point clouds, meshes).- Image Rectification⠀• First-authored paper accepted for ICASSP 2020 (Self-supervised learning for Fisheye Image Rectification)⠀• Worked on fisheye-lens distortion correction using unpaired image data.⠀• Designed an unsupervised framework with Generative Adversarial Networks backbone to predict image flow maps for fisheye image rectification that greatly outperforms current unpaired image translation methods. Show less

• Lead a team of 30+ students to maintain servers for the Electrical Engineering Department. • Collaborated with different departments in the Student’s Association to work on a variety of different projects such as browser games and websites that support campus-wide marketing campaigns.• Organized weekly lectures and workshops on web development and programming for underclassmen.

• Designed high-performance trading algorithms (alphas) using price-volume, fundamental, and news data.• Won the best price-volume alpha award (awarded to the intern who designed the most creative and rewarding alpha based on price-volume data).