PALAIT / 팔아잇
A sourcing-to-listing workflow for Korean sellers, connecting a web application and Chrome extension.
Project page · PALAIT / 팔아잇 ↗Hi, I’m
Building with care.
Thinking with curiosity.
Software engineer & CS master’s student
Yonsei University · Mobile AI & Systems Lab

A little logic.
A little imagination.
An engineer’s mind.
A researcher’s curiosity.
I’m Eddie (Junyoung Sohn), a product engineer building web and mobile products and developer tools. I’m an M.S. student in Computer Science at Yonsei University’s Mobile AI and Systems Lab, interested in on-device AI and systems.
My product work spans TypeScript, React, Next.js, React Native, and workflow automation. I also build Python developer tools and study systems through C and Unix projects. I develop PALAIT through Nanocraft, finding research questions in the behavior and constraints of software people use.
Between what works
and what’s possible,
there’s a lot to explore.
Web products, open-source tools, and a personal prototype. Each project includes my focus and its current stage.
A sourcing-to-listing workflow for Korean sellers, connecting a web application and Chrome extension.
Project page · PALAIT / 팔아잇 ↗Validation for Python dictionaries and configuration, using ordinary types and no required runtime dependencies.
Project page · zodify ↗Korean address information and search, with attention to address-data relationships and useful page structure.
Project page · THEJUSO / 더주소 ↗A webcam-based posture-sensing prototype using local inference. A practical setting for questions about continuous vision and resource use.
Project page · Posely ↗I’m an M.S. student in Computer Science at Yonsei University’s Mobile AI and Systems Lab (MAIS), advised by Professor Seonghoon Park. My interests include on-device AI, inference latency, and resource use in continuously running applications.
Intelligent applications that run on the devices we use every day.
Learning about latency and resource constraints in continuously running applications.
Connecting independent prototype work to questions worth investigating.
Independent projects are presented separately from my lab affiliation. Research outputs and publications will be added when ready to share.
Start with a real need. Let the complexity earn its place.
A clear sentence, a considered interaction, a system that behaves as expected.
Ask better questions. Make the evidence visible. Leave room to change your mind.
On-device AI and systems. Advised by Professor Seonghoon Park.
PALAIT, zodify, THEJUSO, and Posely.
React Native applications, Next.js admin tools, and workflow automation.
A question, an idea, or a different way of looking at things.
I’d love to hear it.