Jungsik Choi, Ph.D.
Staff Researcher
AI System Platform
Samsung Advanced Institute of Technology (SAIT)
Samsung Electronics
🎓 Biography
My name is Jungsik Choi (Korean: 최정식). I am a staff researcher at the AI System Platform of Samsung Advanced Institute of Technology (SAIT), where I focus on designing next-generation computer systems to enable and accelerate AI services. To build world-class AI systems, I take a holistic approach across the entire computing stack—from microarchitecture and operating systems to compilers, system software, AI frameworks, and applications. 🚀
Before joining SAIT, I worked as a postdoctoral researcher at the Electronics and Telecommunications Research Institute (ETRI) and Sungkyunkwan University (SKKU). I completed my Ph.D. in Electrical, Electronic, and Computer Engineering at SKKU in 2019 under the guidance and support of Prof. Hwansoo Han. You can find more about my work and publications online.
📖 Publications
Papers
- Libpubl: exploiting persistent user buffers as logs for write atomicity, HotStorage ‘21
- Libnvmmio: Reconstructing Software IO Path with Failure-Atomic Memory-Mapped Interface, USENIX ATC ‘20
- Multithreaded double queuing for balanced CPU-GPU memory copying, SAC ‘19
- Efficient Memory Mapped File I/O for In-Memory File Systems, HotStorage ‘17
- In-memory file system with efficient swap support for mobile smart devices, IEEE Transactions on Consumer Electronics 62
- Accelerating loops for coarse grained reconfigurable architectures using instruction extensions, ACM RACS ‘11
Patents
- Method and apparatus with data preloading, US19/202768
- Device and method with reduction of cache line fetches, US19/259074
- Device and method with single-level page table for obtaining physical addresses, US18/945080
- Electronic device and method with tensor management and prefetching, US18/776,471
- Electronic device and method with efficient memory management, US18/776,471
- Method and apparatus with memory management based on memory group, US12,417,174
- Apparatus and method for adaptive checkpoint in intermittent computing, US12,197,286
- Method of handling page fault in nonvolatile main memory system, US10,719,263
- Memory mapping management method for nonvolatile main memory system, US10,185,654
- Memory mapping method of nonvolatile memory system and system for providing the memory mapping method, US9,501,424
📜 Certifications
- System Expert, Seoul National University (SNU), Nov. 2025
🏆 Awards
- Best Presentation Award, KIISE, Dec. 2015
- Best Paper Award, ACM RACS, Nov. 2011