About
I’m currently at OctoML where I work as a senior software engineer in the MLSys team on inference, optimizations and hardware acceleration for machine learning models. I also contribute to Apache TVM & PyTorch on stuff like model compilation, coverage and auto-tuning.
I was a software engineer at Snowflake working on Data Lake and semi/un-structured data storage, like parquet and delta lake. I also interned at Microsoft Research Asia, working on transactional distributed databases.
I earned my master’s degree from UIUC CS and bachelor’s degree from PolyU. I worked with Prof. Vikram Adve on compilation of Deep Learning models with HPVM/LLVM during my master’s thesis and Prof. Zili Shao on key-value store.
Recent Work
-
An optimizing compiler for ONNX models on heterogeneous systems
Yuanjing Shi, Advisor: Vikram Adve
Dept. of Computer Science, University of Illinois at Urbana-Champaign (M.S. Thesis), 2020.
-
Leyenda: An Adaptive, Hybrid Sorting Algorithm for Large Scale Data with Limited Memory
Yuanjing Shi, Zhaoxing Li
Finalist, ACM SIGMOD Programming Contest (SIGMOD’ 19), 2019.
-
SQLiteKV: An Efficient LSM-tree-based Lightweight Database Engine for Mobile Devices - (Extended version available at IEEE TCAD 2018)
Yuanjing Shi, Zhaoyan Shen, Zili Shao
23rd Asia and South Pacific Design Automation Conference (ASP-DAC), 2018.
Misc
- Contact me at yj at shingjan dot com
- My primary blog