CV
Education, research, competitions, and skills.
Contact Information
| Name | MinJae Choi |
| Professional Title | Undergraduate Researcher — Efficient ML & Foundation Model Systems |
| minjaechoi.cs@gmail.com |
Professional Summary
My research interests lie in Efficient ML, especially for AI foundation models, with a focus on end-to-end foundation model systems spanning architecture design, large-scale training, low-bit quantization, efficient deployment, and inference kernel optimization. I am also interested in theoretical computer science, particularly algorithmic graph structure theory and algorithm design for NP-hard problems.
Experience
-
2026 - Present KAIST, Daejeon
Undergraduate Researcher
Flexible Machine Learning Lab, KAIST
Advisor: Prof. Insu Han.
- Co-advised by Prof. Sukmin Yun (Vision AI Lab, Hanyang Univ. ERICA)
- Collaborated with Samsung Electronics DS Division
-
2026 - 2026 Daejeon, South Korea
Head of Promotion
KAIST Include
-
2025 - 2025 KAIST, Daejeon
Student Ambassador
Google Gemini Student Ambassador — Team KAIST Include
- Received Phase 1 Second Best Activity Team Award
Education
-
2025 - Present Daejeon, South Korea
B.S.
Korea Advanced Institute of Science and Technology (KAIST)
Computer Science; AI Systems (Double Major); Mathematics (Minor)
- Advisor: Prof. Sungeui Yoon (School of Computing)
Skills
Training
Projects
-
2026 - Present Samsung Electronics (Device Solution Division) — Trillion-Scale Language Model Quantization
Industrial collaboration project on near-1-bit quantization for trillion-scale foundation models, carried out with the Flexible Machine Learning Lab.
-
2026 - 2026 ZipJiGi — 2026 KB Financial AI Challenge
Hosted by KB Kookmin Bank, with Yelim Lee.
- 🏆 Won Excellence Award (KT Award), Top 1.33% — $2,207.86 cash prize, hiring perks
- GitHub repo
- Demo video
-
2026 - 2026 NVIDIA Nemotron 3 30B Model LoRA Adapter Development
2026 NVIDIA Nemotron3 Reasoning Challenge, with Jiseok Son and Seungsu Lee.
-
2026 - 2026 LG AI Research EXAONE 4 1.2B Model Quantization
2026 LG AI Research EXAONE Compression Competition, with Jiseok Son, Seungsu Lee, and Chaewon Lee.
- 🏆 Top 3% in Semifinal Round, advanced to the Final Round
-
2025 - 2026 AMPLIFY — Question Driven Collective Learning Platform
Participated in the Microsoft Imagine Cup 2026 MVP Round.
-
2022 - 2023 EASY COACH — Vision-based Coaching Platform
A computer-vision-based posture correction tool that automatically analyzes and coaches exercise form.
Certificates
- SQL Developer (SQLD) - Korea Data Agency (K-DATA) (2025)
Languages
Advisor
References
- Yelim Lee
BS in Software Engineering and Financial Engineering at Ajou University
- Jiseok Son
BS in Computer Engineering at Hongik University
- Seungsu Lee
BS in Computer Engineering at Hansung University
- Chaewon Lee
BS in Mathematics and Computer Science at Sogang University