Junhyeok Park

About Me

An AI & Software Engineer dedicated to creating better user experiences with technology.

With a passion for building high-performance, scalable, and maintainable systems, I am deeply interested in developing AI models that understand context and adapt to dynamic user interactions. By consistently exploring advanced methodologies and integrating them into real-world applications, I strive to bridge the gap between complex algorithms and practical utility. These efforts to solve systemic challenges have naturally led to meaningful technical outcomes, including patents and academic publications.

I believe in the synergy of growing 'together' rather than alone.

As a strong believer in continuous improvement, I enjoy working in collaborative environments where diverse ideas intersect. I actively share my knowledge of AI and programming with colleagues, finding optimal solutions through open discussion. Based on my experience tackling multi-disciplinary problems alongside diverse team members, I aim to contribute to our team's technical growth and help foster a positive, impactful development culture.

Education

M.S. in Metaverse, KAIST (Aug. 2022 - Aug. 2025)
Advisor: Prof. Woontack Woo
Thesis: Sequential Context Engineering for Zero-Shot Recognition of Procedural Tasks in Egocentric AR

B.S. in Artificial Intelligence, Sungkyunkwan University (Mar. 2018 - Feb. 2022)
Advisor: Prof. Choong-Wan Woo
Capstone Project: Building a Virtual Environment for an AI Agent with Interoception

Experiences

Graduate Researcher, KAIST (Aug. 2022 - Aug. 2025)
- Pioneered research on sequential context engineering for zero-shot recognition in egocentric AR — M.S. Thesis [3].
- Engineered a holistic state tracking system and integrated quantified-self data pipelines for personalized AR interactions [1][5][6].
- Built an edge computing motion detector for Tetris using TensorFlow Lite and Arduino — Term Project, Most Creative Award

Visiting Research Scholar, Carnegie Mellon University (Aug. 2023 - Feb. 2024)
Designed robust fine-tuning paradigms for vision-language models under distribution shifts [4].

Undergraduate / Post-bacc Researcher, IBS Center for Neuroscience Imaging Research (Mar. 2018 - Aug. 2022)
- Built reinforcement learning models and a 3D virtual environment to test AI agents with interoceptive abilities — Capstone Project, Best Poster Award [2].
- Designed a neural network to detect abnormal electrocardiogram signals, resulting in a registered patent [7].

Publications

1. Eunhwa Song, Taewook Ha, Junhyeok Park, Hyunjin Lee, Woontack Woo. "Holistic quantified-self for context-aware wearable augmented reality". International Journal of Human-Computer Studies, 2025.
2. Sungwoo Lee, Jungmin Lee, Kim Sohee, Hyebhin Yoon, Shinwon Park, Junhyeok Park, Jaehyuk Bae, Seok-Jun Hong, Choong-Wan Woo. "EVAAA: A Virtual Environment Platform for Essential Variables in Autonomous and Adaptive Agents". NeurIPS 2025 Datasets and Benchmarks Track.
3. Junhyeok Park, Woontack Woo. "Sequential Context Engineering for Zero-Shot Recognition of Procedural Tasks in Egocentric AR". ISMAR-Adjunct, 2025.
4. Changdae Oh, Mijoo Kim, Hyesu Lim, Junhyeok Park, Euiseog Jeong, Zhi-Qi Cheng, Kyungwoo Song. "Towards calibrated robust fine-tuning of vision-language models". NeurIPS 2023 Workshop on Distribution Shifts.
5. Eunhwa Song, Taewook Ha, Junhyeok Park, Hyunjin Lee, Woontack Woo. "Holistic Quantified-Self: An Integrated User Model for AR Glass Users". ISMAR-Adjunct, 2023.

Patents & Copyrights

6. Face Detection and Keypoint Mapping Based on HoloLens2. (KR Registered, 2024)
7. Method for Judging Abnormal Electrocardiogram using Artificial Neural Network. (KR Filed, 2021)