Hi! I’m Yuchen Tang, a Member of Technical Staff at Everpure (formerly Pure Storage), working on the FlashBlade and FlashArray storage systems. I focus on frontier dataplane technologies, spanning offloading techniques, such as NVIDIA Data Path Accelerator (DPA), and cache injection technologies, such as Intel Data Direct I/O (DDIO) and AMD Smart Data Cache Injection (SDCI).
Before graduate school, I spent two years at Huawei as an Operating System Engineer, working on tiered memory systems and HPC systems.
Interests
- Low-level networking stack: Mellanox CX series, Broadcom Thor series
- Kernel bypassing techniques: DPDK, RDMA
- Compute Express Link (CXL), in both industry and research: memory expansion with CXL 1.1 devices at Huawei, and VM live migration over a CXL 2.0 memory pool at Stony Brook.
- GPU: worklog of crafting a CUDA SGEMM kernel that reaches 80–85% of cuBLAS, and how I keep optimizing it.
Education
- 2024 – 2025 M.S. in Computer Engineering, Stony Brook University
- Research with Prof. Michael Ferdman and Prof. Erez Zadok
- 2019 – 2022 B.S. in Computer Science, University of Michigan
- 2017 – 2019 B.S. in Electrical and Computer Engineering, Shanghai Jiao Tong University
More details are in my resume. Feel free to reach out if you’d like to connect!