About Me
I received my PhD from the Department of Intelligence and Information at Seoul National University in September 2025, advised by Professor Jung Ho Ahn. I am currently a Software Engineer at CryptoLab Inc., where I lead GPU code optimization for homomorphic encryption libraries.
My research interests lie in GPU acceleration and the design of application-specific accelerators informed by profiling and performance analysis of conventional computing platforms. I primarily focus on accelerating cryptographic primitives, including homomorphic encryption, private information retrieval, and zero-knowledge proofs.
I have authored numerous publications in top-tier computer architecture venues (ISCA, HPCA, MICRO), with a cumulative citation count exceeding 1,000.
Research Interests
Hardware ↔ Algorithm Co-Evolution
I believe the best performance breakthroughs come from understanding both sides of the stack. My research pursues two tightly coupled directions:
Profiling algorithmic workloads to uncover bottlenecks, then designing accelerator architectures and GPU optimizations tailored to the computation's unique characteristics.
Leveraging architectural insights to reshape algorithms themselves — restructuring data layouts, reordering operations, and reducing memory footprints so that software and hardware amplify each other.
Computer Architecture
Designing application-specific accelerators through systematic profiling and performance analysis of conventional computing platforms. Focus on memory-centric optimization and efficient hardware utilization.
Homomorphic Encryption
Accelerating fully homomorphic encryption (FHE) operations on GPUs and custom ASICs. Proposed BTS, the first bootstrappable FHE accelerator (ISCA 2022, 246 citations).
Privacy-Preserving Computing
Enabling practical privacy-preserving machine learning and private information retrieval through HW/SW co-design. Bridging the gap between theoretical security and real-world deployment.
GPU Acceleration
Hardware-aware optimization of cryptographic workloads on modern GPU microarchitectures. Achieved over 100x speedup for FHE bootstrapping through memory-centric GPU optimization.
Research Focus: From Algorithm to Silicon
Publications
Total citations: 1,000+ | Google Scholar Profile
Selected Publications (First Author)
IVE: An Accelerator for Single-Server Private Information Retrieval Using Versatile Processing Elements
Introduced techniques to solve memory bandwidth bottlenecks in PIR and proposed a tailored accelerator achieving over three orders of magnitude better throughput than prior hardware solutions.
BTS: An Accelerator for Bootstrappable Fully Homomorphic Encryption
Proposed memory access time-driven design for FHE acceleration. Finalist for Intel Hardware Security Academic Awards 2024.
246 citationsAccelerating Number Theoretic Transformations for Bootstrappable Homomorphic Encryption on GPUs
Analyzed computational traits of NTT in HE context. The proposed implementation influenced subsequent GPU-acceleration works.
103 citationsAll Publications
A Five-Year Journey to Accelerate Homomorphic Encryption with GPUs, Demonstrated by Sub-25ms CNN Inference
Accepted (to appear)Anaheim: Architecture and Algorithms for Processing Fully Homomorphic Encryption in Memory
8 citationsBiography
Software Engineer
CryptoLab Inc.
Responsible for hardware-aware optimization of the company's latest GPU library, HEaaN2. Profiling existing code to identify major bottlenecks and proposing effective solutions. Leading GPU code optimization efforts for newly adopted algorithms in homomorphic encryption.
Ph.D. in Intelligence and Information
Seoul National University (SNU)
Advisor: Professor Jung Ho Ahn
GPA: 3.97 / 4.3
Focus: Computer architecture, Homomorphic encryption acceleration, HW/SW co-design
B.S. in Electrical and Computer Engineering
Seoul National University (SNU)
GPA: 3.6 / 4.3