Jonghyun Choe
Hi, I am a Masters in Computer Science (MSCS) student at Carnegie Mellon University. I am interested in optimizing performance of large-scale machine learning systems, from perspectives of compiler, scheduling, and kernel design.
I am currently working on ML compilers and kernels at CMU Catalyst Lab, under the guidance of Professor Tianqi Chen.
Having lived in the United States and China for 10 years, I enjoy learning new things and connecting with people. Feel free to reach out if you’d like to chat!
News
| Jan 12, 2026 | I joined Professor Tianqi Chen’s lab to work on research in ML compilers and efficient systems. |
|---|---|
| Nov 16, 2025 | I was invited to attend xAI Tech Day in New York and had the opportunity to interview with co-founder Greg Yang. Following our conversation, I was connected with the xAI inference team. |
| Jul 7, 2025 | I was selected to participate in Jane Street SEE, and had the opportunity to meet software engineers and learn about its systems. |
| Mar 15, 2025 | Our paper “Worst Case Response Time Analysis for Completely Fair Scheduling in Linux Systems” is accepted to Real-Time Systems Journal! |
| Feb 20, 2025 | I had the opportunity to receieve mentoring from Google Korea CEO Harrison Kim and discuss career trajectories related to ML systems. |
| Feb 5, 2025 | I started participating in AI Compiler Study, led by Taehoon Kim. Our goal is to study reverse engineering of DeepSeek-V3 and implement optimization techniques. |
| Jan 22, 2025 | I won the Grand Prize at 2025 AWS Solve-A-Thon! Our project aimed to improve reasoning ability of younger generations through guidance generated with LLM. |
| Sep 10, 2024 | I attended IEEE ETFA 2024 held in Padova, Italy with the generous funding from Specialized Semiconductor Scholarship. |
| Mar 15, 2024 | I received Outstanding Teaching Assistant Award for Computer Architecture course. |
| Feb 24, 2024 | I won 1st place at Accelerator Programming School CUDA Competition with Jake Hyun. |
About
Education
-
2025.08 - Present -
2019.03 - 2025.08
Awards
- 2024.02.01
1st Place, Accelerator Programming CUDA Competition
Next Generation Supercomputing Center
Achieved 50,000 times speedup in model inference throughput through CUDA optimizations
- 2024.02.01
Full Merit Scholarship
Seoul National University
Full tuition awarded for outstanding academic performance
- 2017.12.01
USA Mathematical Olympiad Qualifier
American Mathematics Competitions
Top 250 out of 500,000 participants in the U.S. math competition
Publications
Work
- 2025.01 - 2025.05
Research Intern
NVIDIA / Speech AI Lab
Advisor: Namjoon Kim, NVIDIA / Speech AI Lab, Hyungon Ryu, NVIDIA
- Optimized LLM inference on edge devices by exploiting Jetson’s unified memory architecture
- Researched KV Cache reuse system with external storage devices including SSDs
- 2024.01 - 2024.10
Research Intern
Architecture and Code Optimization Lab
Advisor: Prof. Jae W. Lee, SNU
- Researched CPU architecture for hiding memory stall latency through software-defined context
- Researched dynamic low-rank approximation to improve LLM inference
- 2023.07 - 2023.08
Research Intern
Co-design and Parallel Processing Lab
Advisor: Prof. Soonhoi Ha, SNU
- Developed Linux scheduler simulator for schedulability analysis
- Researched probabilistic schedulability analysis for more efficient task scheduling and higher system utilization
- 2019.12 - 2020.08
Research Intern
Big Data Studies Lab
Advisor: Prof. Javier Cha, SNU
- Researched energy efficiency of data centers, evaluating hardware, power delivery, and cooling components
- Analyzed physical limitations of storage devices affecting big data operations
Languages
| Korean | |
| Native |
| English | |
| Fluent |