Hi! I am currently working as a Cryptography Engineer at CryptoLab Inc, Seoul, Korea. I recently earned
my PhD at Seoul National University (SNU), Department of Mathematical
Sciences.
My PhD advisor was Prof. Jung Hee Cheon.
Here is my Curriculum Vitae. I work on cryptography, focusing on (threshold)
fully homomorphic encryption (FHE) and post-quantum cryptography (PQC) mainly focusing on lattice-based
ones. I am also interested in privacy enhancing technologies (PET).
Currently, I am a member of Team SMAUG(-T) and Team
HAETAE, the winners of the Korean Post Quantum Cryptography (KpqC) Competition
for Korean PQC standard. HAETAE was also submitted to NIST PQC Additional
Signature Round 1.
I am presenting an Invited Talk about HAETAE and SMAUG-T, the KpqC Selected Algorithms, at 2025 KMS Spring Meeting in April.
The paper suggesting homomorphic LookUp Table evaluation is accepted to Journal of the Korean Mathematical Society.
HAETAE digital signature scheme and SMAUG-T KEM scheme are selected as final winners of the KpqC competition.
The paper analyzing the KpqC Round2 changes of SMAUG-T KEM is accepted to IEEE ACCESS. The paper introduces new designs to SMAUG-T making it more timing-secure and efficient with less failures.
The paper "Attacks Against the IND-CPAD Security of Exact FHE" is accepted to ACM CCS 2024. I am attending CCS, Salt Lake City, US and presenting at the Doctoral Symposium (Mid Oct. 2024).
Cloud computing enables complex operations that clients with limited resources cannot compute alone. But is it safe? Wouldn't cloud providers use your data for other purposes? While general encryption only securely stores your data, Homorphic Encryption (HE) enables computation on encrypted data. HE enables cloud computing while ensuring data privacy. I am interested in broad topics in HE, e.g., security notions, accelerations, applications, and extensions to the threshold/multi-party setting.
Will cryptosystems still be secure in the future when quantum computers are commercialized? The security of most modern cryptosystems is guaranteed by its base hard problems. However, most of the hard problems currently in use can be easily solved by quantum computers. Post-Quantum (or quantum-resistant) cryptography designs cryptosystems using hard problems that are difficult to solve even with quantum computers. I am particularly interested in cryptosystems based on lattice problems.
Grafting: Decoupled Scale Factors and Modulus in RNS-CKKS,
with Jung Hee Cheon, Minsik Kang, Jaehyung Kim, Seonghak Kim, Johannes Mono, and Taeyeong Noh,
Eprint Archive 2024.
In this paper, we revisit the arithmetic structure of CKKS FHE scheme and introduce Grafting, a technique that decouples the scale factor from the modulus. This enables optimized RNS packing and flexible rescaling, while reducing public key sizes and execution time. Our latest experiments show 1.92× faster bootstrapping, 1.89× faster Bit-CKKS bootstrspping, and 62% smaller public keys.
This work is part of my doctoral dissertation.
Attacks Against the IND-CPAD Security of Exact FHE,
with Jung Hee Cheon, Alain Passelègue, Damien Stehlé, and Elias Suvanto,
ACM CCS 2024.
In this paper, we address the IND-CPAD insecurity of exact FHE schemes, which was previously believed to apply only to approximate FHEs. We identify vulnerabilities arising from non-negligible incorrectness, where the probability of incorrect evaluation is not exponentially small. Our attacks efficiently recover secret keys from BFV/BGV, DM/CGGI, and the exact variant of CKKS, exposing vulnerabilities in current parameter choices.
This paper received the Grand prize at the Korean National Cryptography Contest in 2024.
HAETAE and SMAUG-T,
Final Winners of KpqC Standardization Competition.
After 3 years of tremendous efforts—public evaluations, numerous improvements, and security validations—HAETAE and SMAUG-T were selected as the final winners of the KpqC Competition in January 2025. The competition aims to select Quantum-secure Digital Signatures and KEM candidates for Korean Standards, similar to NIST's PQC standardization process. These schemes will, in the near future, become part of the Korean Standard (KS). As a main contributor to both schemes, I am truly proud of the progress we've made.
