- 🔭 I am working on database and LLM system.
- 💻 MPhil in DGGroup@SUSTech.
- 👨🎓 Graduated from BUPT in
2022
. - 🔙 I am happy with my internships in the following places.
2019
CASIA2021
Alibaba Ant Group2022
Huawei GaussDB Group2024
RS3Lab@EPFL2025
Systems@TUDa
- 📫 yangshen.d@outlook.com
- 🏠 dengyangshen.netlify.app
🌋
Working on Data x AI
🚀Master student in DBGroup@SUSTech.
- Shenzhen, China
-
04:16
(UTC +08:00) - https://dengyangshen.netlify.app
- https://orcid.org/0009-0003-9487-1455
Highlights
Pinned Loading
-
DBGroup-SUSTech/GPU-Merkle-Patricia-Trie
DBGroup-SUSTech/GPU-Merkle-Patricia-Trie PublicAn optimized Merkle Patricia Trie implementation on GPU, fully compatible with and integrable into Ethereum. The paper is published on VLDB 2024.
Cuda 12
-
DBGroup-SUSTech/GPUDB-Prefetch
DBGroup-SUSTech/GPUDB-Prefetch PublicSource code of our DaMoN@SIGMOD 2024 paper "How Does Software Prefetching Work on GPU Query Processing?"
Cuda 9
-
vllm-project/vllm
vllm-project/vllm PublicA high-throughput and memory-efficient inference and serving engine for LLMs
Something went wrong, please refresh the page to try again.
If the problem persists, check the GitHub status page or contact support.
If the problem persists, check the GitHub status page or contact support.