Information everywhere, but no overview
Every text, paper, and email uses a different format — users re-read, copy, and double-check.
Human-Centered AI · Product · Engineering
I turn everyday complexity
into clear, usable
AI experiences.
Scattered expense records, confusing guitar tone settings — I'm Lina Noh, and I turn the complexity people face every day into clear flows and working products.
Product Strategy·UX & IA·Software Prototyping·Applied AI
Problem
Technology often pushes the organizing
and configuring onto users.
Powerful tools, but too hard to use
Multi-effects units offer endless menus, and the path to the sound you want feels impossible.
AI without reasons is hard to trust
Good automation provides sources, explanations, corrections, and undo — not just results.
“Technology should help people understand and choose,
not demand more studying and organizing.”
Guide
I listen to complex situations, structure them,
and make them immediately usable.
- ObserveFind recurring friction in real life, work, and music-making contexts.
- StructureDefine one core problem and its success criteria.
- JudgeCompare alternatives and publish the reasons and limits of each choice.
- BuildPrototype and validate with real inputs.
- ExtendGrow the experience into Human-Centered AI research questions.
Process
A four-step working plan
STEP 01
Understand
Understand users, context, gear, and constraints.
STEP 02
Frame
Define the core problem and the roles of AI and users.
STEP 03
Build
Build the smallest working flow and validate it.
STEP 04
Learn
Design the next version from errors and feedback.
Featured Projects
Two deep case studies
One philosophy in two different domains: AI organizes the complexity, while users stay able to understand and correct the results.
Scattered statements, one traceable ledger.
AI reads and structures statements arriving by text, paper, and email into a monthly ledger. Users verify against the source, then edit and confirm.
- UsersPeople who manage their own living expenses
- AIExtraction, structuring, duplicate candidates
- PeopleVerify sources, edit, final confirmation
7월 지출
2,149,180원
이번 달 요약
쇼핑1,494,380원
생활600,000원
식비30,900원
✎ 직접 적기📷 영수증 찍기
최근 내역
쿠팡(주) 쇼핑 · 7.27-7,400원
테스트상점 기타 · 7.26-2,000원
월세 생활 · 7.26-600,000원
Focus on playing. TonePilot handles the starting point.
Based on your gear and desired sound, it proposes a starting multi-effects setup — and explains "why this value" in a player's language.
- UsersGuitarists who find multi-effects units overwhelming
- AIChain & parameter starting points
- PeoplePlay, give feedback, adjust, save
후크 (예술이야 반복)1:40 → 1:57
짧고 강렬한 후크 파트, 하이게인 유지하며 딜레이/리버브 강조.
AMP · Gain7/10하이게인 유지
DLY · Dual DelayTime 450ms리듬 악센트 강조
REV · SwellDecay 40%악센트마다 스웰감
👉 P04 프리셋 호출, 픽업 1번 포지션 유지.
Research
I want to research AI that helps people
understand and decide better.
Personal Information & Everyday AI
How can AI organize personal documents while preserving sources, privacy, and the right to correct?
Human-AI Co-Creation for Music
How can AI propose tone starting points while keeping the player's taste and agency?
Explainable Applied ML
How can predictions from baseball and K-pop data become understandable explanations?
About
From the field to products, from products to research.
Skills
A capability map linked to evidence, not a list of logos.
Writing
Notes on product judgment, failures, and data problems.
Let's Talk