SandAI
Sep 2024 – Present
- Creator of MagiCompiler and Athena, SandAI's internal inference engine.
- Core contributor to MAGI-1
About Me
I work at SandAI, where I created MagiCompiler and Athena (SandAI's internal inference engine), and serve as a core contributor to MAGI-1.
My research centers on AI for Systems, spanning AI compilers, inference optimization, and deep learning infrastructure. I aim to make AI more accessible, efficient, and impactful—creating innovations that democratize AI and unlock its potential for everyone.
I received my M.S. from Tsinghua University (2022) and B.S. from Dalian University of Technology (2019).
Experience
Baidu – PaddlePaddle
Jul 2022 – Aug 2024
- Top contributor to PaddlePaddle on GitHub; deeply involved in the AI compiler, custom operator mechanism, and other core modules.
Advertising Algorithm Intern Alibaba – Alimama Jul – Sep 2021
Search Algorithm Intern Microsoft Research Asia Mar – Jun 2021
Recommendation Algorithm Intern Tencent – WeChat Reading Apr – Aug 2020
Education
M.S., Software Engineering
2019 – 2022
- Glodon First-Class Scholarship, 2020 (Top 10%)
- KDD Cup AutoML Track, 4th Place, 2019
B.S., Software Engineering (dual degree in Japanese)
2015 – 2019
Dalian University of Technology
- National Scholarship, 2015 (Top 1%)
- Qu Bochuan Scholarship (屈伯川奖学金), 2016 — university-wide special award, 10 recipients
Publications
- H. Jia, Y. Jiang, C. Zhong, et al. “TTDeep: Time-Triggered Scheduling for Real-Time Ethernet via Deep Reinforcement Learning.” IEEE Global Communications Conference (GLOBECOM), 2021. [pdf]
- J. Sun, H. Jia, B. Hu, et al. “Speeding Up Very Fast Decision Tree with Low Computational Cost.” International Joint Conference on Artificial Intelligence (IJCAI), 2021. [pdf]
- C. Zhong, H. Jia, H. Wan, et al. “DRLS: A Deep Reinforcement Learning Based Scheduler for Time-Triggered Ethernet.” International Conference on Computer Communications and Networks (ICCCN), 2021. [pdf]