Focus Areas

  • GPU-accelerated parallel computing
  • CUDA / HIP / SYCL performance engineering
  • Multi-GPU scalability and data-movement optimization
  • Performance-critical imaging and systems tooling

Photography

Photography complements my systems work through observation, timing, and iteration. I shoot both film and digital, from compact film cameras such as the Contax T2 and Nikon 35Ti to Fujifilm X-series bodies including the X-H2, X-T5, and X100VI.

Hill of Mild Seven
Hill of Mild Seven
The Lonely Christmas Tree
The Lonely Christmas Tree
Tokyo Odaiba Rainbow Fireworks
Tokyo Odaiba Rainbow Fireworks
Spring Whispers
Spring Whispers
Sunset Eyes, Eternal Soul
Sunset Eyes, Eternal Soul
Flickr500px

Selected public work

PyTurboJPEG

Python native wrapper

High-throughput Python wrapper for libjpeg-turbo across x86 and ARM platforms.

Pythonlibjpeg-turboSIMD

SyclReproSum

Reproducible SYCL library

Header-only library for fast, bit-reproducible floating-point summation across CPUs, GPUs, vendors, and SYCL backends.

SYCLC++17reproducible numerics

bokeh.cc

Image workflow app

Browser image viewer for local folder review, comparison, and selected export.

browser APIsimage workflowbatch export

exif.pl

Photo metadata app

Client-side EXIF metadata viewer. Designed to work well on mobile devices.

Exif metadatamobile devices friendly

Education

National Tsing Hua UniversityMS, Computer Science