Curriculum vitaeUpdated October 2026

Yunmin Cha 차윤민

mrcha033@yonsei.ac.kr · mrcha033.github.io · github.com/mrcha033 · linkedin.com/in/ynmncha

Systems research, engineering, and university service.

Education

Yonsei University

March 2022 – present

Seoul, Republic of Korea

B.B.A. candidate in Business Administration; second major in Computer Science.

Undergraduate–Master’s Integrated Track in Computer Science, School of Computing.

Selected coursework

Operating Systems · Multicore and GPU Programming · Computer Systems · Computer Architecture · Data Structures · Decision Under Uncertainty · AI and Operations Analytics · Linear Algebra

Research

Systems for predictable and efficient LLM inference, and decisions under uncertainty.

Sharing GPUs among LLM workloads

July 2026 – present

S3 Lab, Yonsei University · Advisor: Prof. Jinkyu Jeong

Scheduling beneath LLM serving engines: dividing a GPU among workloads whose demand is uncertain.

My part: the problem formulation, the scheduling policies, and a native serving engine.

Causal performance analysis for GPU programs

July 2026 – present

S3 Lab, Yonsei University · Advisor: Prof. Jinkyu Jeong

Estimating what speeding up one region of GPU code would do for the whole application.

My part: the profiler, the binary instrumentation beneath it, and the experiments that check its predictions against real optimizations.

Moving data from storage to GPU memory

July 2026 – present

S3 Lab, Yonsei University · Advisor: Prof. Jinkyu Jeong

The design space of direct storage-to-GPU I/O, and which processor should control it.

My part: the Linux driver and GPU library the comparisons run on, and the measurements.

AI and ordering decisions

October 2026 – present

When AI that reads documents and asks managers questions reduces the loss of an ordering decision.

My part: the decision model and its AI system, the pilot experiments, and the software for the study with participants.

Open source

SGLang upstream pull requests

July 2026

A fix that lets the copy-engine-free transfer handle any expert count up to 512, and a consecutive-prefill budget that limits decode starvation. Both open as of October 2026.

hw-codesign

2026

An agentic PCB design and validation toolchain (Python CLI, MCP server, and agent plugin); released on PyPI as v0.1.4.

Honors and scholarships

Top Honors (최우등상)

Fall 2022; Spring 2023

Semester academic honors, Yonsei University.

Honors (우수상)

Spring 2026

Semester academic honors, Yonsei University; top 10%.

Industry–Academia Project Excellence Scholarship

Spring 2026

SW-Centered University Program, Yonsei University; KRW 5,000,000.

Industry experience

CTO, PREP

May 2025 – present

Seoul, Republic of Korea

Led architecture and delivery of a production platform serving real users, including backend APIs, relational data models, deployment automation, and operational tooling.

Owned reliability and latency work across releases through CI/CD, structured logging, health monitoring, and repeatable incident-debugging workflows.

Leadership and university service

Yonsei University Student Council

2025

Chair, Audit Committee · Member, Election Management Committee

Yonsei University Student Clubs Union

2024 – 2026
  • Interim Committee Chair, Religious Clubs Division2026
  • Member, General Operations Office2026
  • Development Team Lead, Pub Task Force; Member, Audit Committee2025, 2026
  • Chair, Religious Clubs Division2025
  • Deputy Director, Club Representatives’ Workshop2025
  • Vice Chair, Religious Clubs Division2024

Skills and languages

Programming
Python, C, C++, CUDA C++, Bash, TypeScript/JavaScript, SQL
GPU and ML systems
CUDA, SASS instrumentation, NVBit, CUPTI, Nsight Systems/Compute, NVML, cuBLASLt, cuDNN, NCCL, vLLM, SGLang, PyTorch
Systems and I/O
Linux kernel modules, NVMe, asynchronous I/O, CMake, Docker, PostgreSQL, SQLite
Research infrastructure
Pytest, Git, GitHub Actions, LaTeX, deterministic experiments, provenance and telemetry instrumentation
Languages
Korean (native), English (fluent), Spanish (intermediate)