Kaiyi Zhao

Software Engineer Intern

National Institutes of Health (NIH)•  November 2023 - August 2024•  Remote

By accelerating and optimizing Flye genome assemblers, particularly during the polishing stage, achieved a more than fourfold speedup on a bacterial dataset and a twofold speedup on the human genome dataset. Introduced a more efficient multithreading architecture, significantly reducing mutex contention by minimizing mutex lock time and acquisition frequency. Applied Advanced Vector Extensions to process multiple reads simultaneously, effectively parallelizing the polishing process.

Software Engineer Intern

SenseTime•  November 2021 - February 2022•  Beijing

Developed and optimized perception algorithms for SenseTime's autonomous driving solutions, resulting in deployment across 1.95 million vehicles over 90 models. Deployed deep learning networks to SenseTime's intelligent driving vehicles. Rewrote preprocessing and post-processing modules using C++ and CUDA. Utilized TensorRT to convert the main network component for efficient in-vehicle system operation.