News
- 2026.06 Reorganizing this site into an academic profile and learning archive.
- 2024.03 Added notes on problem analysis
- 2024.02 Added notes on fine-tuning LLMs
About Me
Hello. My name is Junbo Koh, and I am a master's student in Educational Technology at Seoul National University.
I use this site as a learning archive for ideas I want to keep revisiting: software development, learning experience design, human-computer interaction, and generative AI. The notes here are intentionally practical. They help me connect what I study with what I build, teach, and evaluate.
My current focus is understanding how people learn with emerging technologies, and how AI-supported tools can help learners reason, practice, reflect, and collaborate more effectively.
Ongoing Research
- Learning experience design: designing learning environments that support practice, feedback, reflection, and transfer.
- Human-computer interaction: understanding how people make sense of interactive systems and where usability breaks down.
- Generative AI and LLMs: exploring how language models can support learning, writing, inquiry, and tool use.
I am interested in the intersection of learning sciences, educational technology, and human-centered AI. More specifically, I want to study how learners interact with intelligent tools, how those tools shape learning processes, and how design decisions can make AI systems more explainable, useful, and accountable in educational contexts.
Publications
Publications will be added as they become available.
Writing / Notes
This site also collects study notes and implementation logs across the topics I return to most often.
- LLM: notes on generative AI, transformer models, prompt engineering, fine-tuning, and RLHF.
- HCI: notes on interaction design, feedback cycles, mental models, task analysis, and design principles.
- Instructional Systems Design: notes on problem analysis, needs assessment, and instructional design foundations.
- Deep Learning for Coders: notes from practical deep learning study with fastai and PyTorch.
Recent Posts
Problem Analysis
This content is based on ‘Foundations of Instructional Design’ book by Seels and Glasgow.
Fine-tuning LLMs
LLM을 파인튜닝하면서 공부한 내용들, 고려해야 할 사항들을 적어본다.
[GenAI] L3-P1-2. Reinforcement Learning from Human Feedback(2)
Proximal Policy Optimization (PPO)
[GenAI] L2-P2-2. Parameter Efficient Fine-Tuning(2)
Prompt Tuning