Lee ChanyangBackend Engineer

Language

← Back to project list

Personal Project

LiveIn Progress

Evidence-Based Engineering Portfolio

An engineering portfolio that began as a quick visual draft and evolved as I identified reading friction, distance from supporting evidence, and public-disclosure boundaries.

Role
Portfolio Designer / Frontend Engineer
Responsibility
Sole responsibility for planning, content modeling, and frontend implementation
Period
2026.07–Present · Operating and Improving
  • Next.js 16
  • React 19
  • TypeScript
  • Claude · Codex

Project Overview

I reframed the portfolio redesign as a publicly shareable case study

Although there was extensive work experience, documentation, retrospective material, and commit history, it lacked an order that readers could assess quickly and clear boundaries around evidence that could be made public.

Unlike the featured backend cases, this personal project did not begin by designing state, authorization, and reprocessing. I built the screens first, then worked backward from friction in the actual reading flow to define requirements.

I handled planning, content modeling, and frontend implementation, separating featured projects into summaries for hiring decisions, detail pages into evidence for technical review, and personal projects into public examples of how I work. Instead of exposing every deliberation, I retained only the choices, revisions, and validation that help readers assess the work.

Ownership

My direct responsibilities

  • Built the initial screens quickly, then introduced structure where reading friction, duplication, and gaps in evidence became visible
  • Set different boundaries to examine first for each project type, without forcing every case into the same sequence as the featured projects
  • Separated source experience into projects, claims, evidence, disclosure boundaries, and limitations, promoting these into content models when needed
  • Restructured the information architecture so featured projects, personal projects, Career/Production Evidence, and the Implementation Timeline provide different levels of detail
  • Iteratively reduced overstated wording and broken anchors through PDRs, persona reviews, link audits, UI and refactoring history, commit and work review records, type checks, and build validation

Constraints

Constraints to address first

  • Company code, actual service names, internal URLs, API paths, database table names, and operational identifiers had to remain private.
  • Recruiters needed to identify strengths quickly, while technical reviewers needed access to design rationale when required.
  • Each project needed a different validation sequence, so I had to avoid implying that one standard flow was mandatory.
  • Working notes, code, documents, and commits could support internal verification, but public pages needed to retain only the claims, evidence, and limitations useful for assessment.
  • With no visitor or conversion data yet, I could report only qualitative persona reviews and link/build validation as outcomes.

Implementation Timeline

I organized the process into a quick draft, screen observation, authoritative evidence, selective exploration, and regression validation

I did not begin with a complete design. The phases describe roughly two weeks from late 2026.07 to early 2026.08, during which I built screens, identified friction, and formalized only the structures that proved necessary. Select a phase to see the problem identified at that stage and how it led to the architecture and verification criteria.

Phase 01 · Reader Needs and Portfolio IA

Reader Needs and Portfolio IA

Phase 01 · Reader Journey

I traced reader questions backward from a quick screen draft

Rather than documenting every reader journey upfront, I built the home and detail pages first, then turned the questions that caused readers to pause into requirements.

Reader Journeys and Home-Page Information HierarchyMermaid

Rendering the diagram.

  1. 01

    Reader Hypotheses

    I began with a hypothesis about who would view the site and how much time they would have, then separated the depth of assessment each reader needed.

    • I defined recruiters as readers seeking positioning, headline metrics, and contact options quickly.
    • I defined technical reviewers as readers seeking problems, constraints, alternatives, limitations, and evidence links.
  2. 02

    Quick Home Draft

    I first built the home, featured-project, and methodology sections, then identified awkward points in the actual reading order.

    • In the initial implementation, I allowed some duplication and lengthy explanations to check whether the overall flow was clear first.
    • After building it, I found that Home, Method, Case Study, Reflection, and CTA each needed a distinct role and level of detail.
  3. 03

    Adjusting the Visual Hierarchy

    After reviewing the screens, I addressed the problem of featured projects, personal projects, and early or supporting experience appearing equally important.

    • I removed early and supporting experience from home-page cards while preserving detail routes and Career/Production Evidence links, so the most representative cases are read first.
    • I limited personal projects to supporting examples of how I analyze requirements and model information, rather than presenting them as production-service outcomes.
    • I made the methodology a navigation path into featured projects rather than a separate explanation.
  4. 04

    Feedback-Based Revisions

    Using persona reviews and PDRs, I made excessive length, distant evidence, and diluted emphasis the criteria for revision.

    • PDR-002, PDR-004, PDR-005, and PDR-008 record reducing featured projects, introducing three levels of information depth, removing filters, and hiding additional experience from the home page.
    • In PDR-009, I cross-checked changes to home, detail, and process interfaces and the reasons for refactoring against commits and work review records.
    • I eased the problem of project details growing into verification documents by introducing a selective Implementation Timeline structure.
  5. 05

    Measurement Deferred

    There are no visitor or conversion metrics yet, so I retain screen review and the unimplemented CTA/PDF work as current limitations.

    • PDR-007 keeps the hiring-conversion CTA and separation of PDF roles in proposed status to avoid overstating completed work.
    • After deployment, project clicks, visits to retrospective links, and email conversions are the next validation targets.

Separating Featured Experience from Career Continuity

Choice

I kept recent core cases as featured projects and made early and supporting experience accessible through Career and evidence links rather than home-page cards.

Problem
Giving every company project the same visual weight diluted the emphasis on the most representative cases.
Rationale
First-time visitors can quickly see the most representative cases and how I work, then explore career continuity through a separate path if needed.

Trade-offEarly projects received less home-page exposure, but the selection criteria for the 3 featured projects and the role of personal projects became clearer.

Project Infrastructure

I operate the portfolio through boundaries for authoritative content, public routes, and validation

I separated three operational boundaries to keep facts consistent and public links and screens intact as portfolio content grows.

Outcome

Verifiable outcomes

  • Organized the portfolio redesign into a publishable case explaining a quick draft, screen observation, evidence auditing, selective navigation, and regression validation
  • Converted backend experience into a layered information structure of home summaries, methodology, project details, and retrospective evidence
  • Normalized 6 project detail pages using the same CaseStudy schema and Implementation Timeline format
  • Displayed only 3 featured projects and personal projects on the home page, preserving early and supporting experience through Career and evidence links
  • Separated decision history from source-resource discovery paths through PDRs, RESOURCE_INDEX, and UI/refactoring/UX audit documents
  • Separated the roles of Portfolio Flow and Implementation Timeline to shorten navigation between shared questions and evidence within projects
  • Explicitly documented limitations around confidential information, unsupported figures, and unimplemented CTA/PDF scope

Limitations

Limitations and unimplemented scope

  • The initial version is Korean-only; English summaries and PDF deliverables are follow-up scope.
  • Visitor, retrospective-click, and email-conversion metrics have not yet been collected, so only qualitative persona reviews and link audits are used as evidence.
  • Company code repositories and internal working notes remain private; the site provides only generalized evidence.
  • Work review records serve as sources for UI and refactoring judgments, while public pages retain only decisions and validation results rather than the original records.
  • The separation of CTA/PDF roles remains proposed in PDR-007 and is not labeled as implemented.