I am focusing on low-latency and high-frequency trading techniques.

At Tencent, I played a key role in developing the Cloud Object Storage system (COS). I am passionate about leveraging the (async) Rust ecosystem in cloud applications, aiming to boost the reliability and performance of cloud infrastructures.

Training data plays a crucial role in the development of machine learning. My primary work at SenseTime involved researching how to accelerate the transfer of training data to GPUs to enhance GPU utilization and reduce costs. Ultimately, we designed and implemented an end-to-end data system, which includes efficient data storage, caching, and a preprocessing module. This significantly improved the training speed (please refer to the DIESEL-related paper below for more details).

PG student course:1. Parallel Programming (MPI, OpenMP, Pthread, CUDA). 2017 (Spring term, Fall term), 2019 Spring term.UG student course:1. COMP4901D - Heterogeneous Parallel Programming. CUDA Programming. 2016 Fall term.2. COMP3311 - Database Management Systems. 2018 Spring term.3. COMP3031 - Principles of Programming Languages. Grammar and Expressions. Flex and Bison. 2019 Fall term.

Performance study on multiple threaded remote direct memory access (RDMA) tech.https://github.com/WANG-lp/RDMA

Accelerating Biological Database Search on Intel Xeon (Phi) Architectures.http://xsw.sdu-hpcl.org/