[SSE3021 Digital Integrated Circuit Design] Schematic and Layout Design of 16x16bit 6T SRAM
This is the report submitted to Digital Integrated Circuit Design course (2026 Spring semester SSE3021, SKKU).
Blog posts
2026
[SSE3021 Digital Integrated Circuit Design] Schematic and Layout Design of 16bit Sklansky Adder
This is the report submitted to Digital Integrated Circuit Design course (2026 Spring semester SSE3021, SKKU).
2025
[Project] CUDA Matrix Multiplication Implementation and Optimization
Implemented and optimized CUDA kernels for scaled dot-product attention, attention linear projection, and Q/K/V projection.
[ISOCC 2025] LLM on FPGA: Squeezing Language Models by Quantization and Multi-Query Attention and Its Efficient Hardware Architecture
Delighted to share that my research has been accepted to an oral session of ISOCC 2025 – 22th International SoC Design Conference! (presented on October 16, 2025)
[tape-out project] quantized BC-ResNet models for keyword spotting and speaker verification
This post details the design and implementation of a quantized BC-ResNet-based neural network accelerator for keyword spotting (KWS) and speaker verification (SV), culminating in a successful ASIC tape-out using Samsung’s 28 nm LPP process. The project demonstrates a hardware–software co-design approach, achieving low power consumption and high accuracy for edge audio intelligence applications.
[project] Senior Year Capstone Design : Front-end and Back-end Design of Digital CNN Accelerator for SubPixel Rendering Algorithm
This project was carried out as a Senior Capstone Design from March to June 2025. It received the Engineering Innovation Prize at the SKKU Senior Capstone Design Contest 2025. I served as the team leader, and my teammates were Somin Kang, Miri Kim, and Jaehwan Shin.
2024
[tape-out project] CNN-based keyword spotting system with framewise incremental computation
This project describes the design and implementation of a CNN-based keyword spotting system, including algorithm selection, RTL design, logic synthesis, and place-and-route for tape-out submission.
2023
[Project] Smart Factory Capstone Design : Rotor Fault Diagnosis Method Based on 1D-CNN LSTM
This paper presents a fault diagnosis method for rotating machinery using a 1D-CNN LSTM architecture. The proposed model effectively captures temporal and frequency-domain features from vibration data, achieving high diagnostic accuracies for various fault types. The study emphasizes the importance of early fault detection in manufacturing environments to prevent downtime and enhance efficiency.