Summary
Yongjin Cho is a research engineer and NLPer with 11 years of experience building high-performance search and production ML systems in South Korea. He helped develop Daum’s core search engine and later improved Kakao’s search and translation infrastructure, launching Kakao’s translation service and its first GPU-based serving system. More recently he led design and operation of a Kubernetes-based GPU ML platform at Upstage, optimizing utilization across 400–500 GPUs and enabling multi-node distributed training with GPU Direct RDMA. His work spans from low-level performance wins—such as SIMD-parallelized index compression—to applied deep learning for MT, ASR, and style-aware translation (Korean honorifics). Holding a Ph.D. in Computer Software Engineering from Seoul National University, he combines rigorous research background with hands-on production expertise across cloud, bare-metal, and hybrid environments. An understated strength is his ability to move models from research into scalable, secure serving systems (including data encryption and billing-aware LLM serving).
11 years of coding experience
13 years of employment as a software developer
Doctor of Philosophy (Ph.D.), Computer Software Engineering, Doctor of Philosophy (Ph.D.), Computer Software Engineering at Seoul National University
Korean, English