Tech Portfolio Best Practices (verified research)

# Tech Portfolio Best Practices (verified research)

AI Summary

Purpose:

• Distilled, adversarially-verified research (2026-06-05 deep-research run:

25 sources, 85 claims extracted, 25 verified, 2 refuted) on how a senior backend/DB/infra engineer's portfolio is evaluated and how to write it.

Key points:

• Recruiters judge in seconds, spend ~5 minutes total → results must be

visible on the first screen. (The popular "7.4 seconds" figure was REFUTED in verification — use the directional claim only.)

• Curate 3–5 deep case studies (documented range 2–10) as a highlight reel;

no credible source endorses listing 30+. "Highlight reel + archive" is the accepted pattern for large bodies of work.

• Korea-specific: #1 hiring criterion is 직무 관련 업무 경험 — 81.6%

(2025 경총 survey of 500 companies; 58.4%→74.6%→81.6% trend).

• Curation filters (priority order): JD alignment → quantifiable impact →

clear personal contribution.

• Write items as STAR / XYZ ("Accomplished X as measured by Y, by doing Z");

34% of hiring managers pass over submissions without measurable results.

• For non-visual backend/DB/infra work, explicitly stated TRADE-OFFS are the

single strongest seniority signal ("what was sacrificed for the choice"), plus architecture diagrams and metric types like latency/cost/error-rate.

• Quantification fallbacks when no hard number exists: honest estimates,

qualitative before/after ("days → hours"), or scope ("50-person team", "수억 개 파일").

Relevant when:

• Writing or reviewing human/portfolio/ items, resumes, or career docs.

Do not read full document unless:

• You need source URLs, refuted claims, or caveat details.

Linked documents:

• ai/worklog/2026/2026-W23.md (application record)
• human/portfolio/index.html

Open Questions

• Verified 2025–2026 portfolio UX trends (card grid + filter, dark theme,

mobile-first) — the research batch surfaced NO verified claims on these.

• Korean recruiter viewing-time behavior vs the US/design-oriented studies.
• Which metric types most impress Korean backend/DB hiring managers

specifically (p99 latency, QPS, replication lag, MTTR, cost-per-query?).

Details

Verified findings (vote = adversarial verification result)

1. First judgment in seconds; whole-portfolio review ~5 minutes; skim focuses

on layout/titles/keywords → lead with results. (3-0 / 2-1 merged. Sources: Ladders eye-tracking via hrdailyadvisor.com, IxDF case-study guide citing Morgane Peng, opendoorscareers.)

1. Korea: 직무 관련 업무 경험 = top criterion, 81.6% (2025 경총, primary

press release verified; jobkorea.co.kr summary). (3-0)

1. Curate 3–5 strong projects (range 2–10); overloading is a named mistake.

(3-0; joshwcomeau.com, nucamp, codingtemple, 8seneca, lovable.dev.)

1. Prioritize by JD 일치도 / 성과 임팩트 / 역할 깊이; structure beats volume.

(3-0; jobkorea, opendoorscareers, Figma recruiter blog.)

1. STAR (상황/과제/행동/결과+수치) shifts duties → impact. (3-0; jobkorea,

enhancv, Jobscan: 58% of recruiters value measurable achievements.)

1. XYZ formula — "Accomplished X as measured by Y, by doing Z" (Laszlo Bock,

primary 2014 LinkedIn article verified). (3-0)

1. Case-study structure: Problem/Hook → Approach → measurable Results

(+ Learnings); 34% of hiring managers pass over no-metrics resumes; 86% visit portfolio links, 71% say it impacts hiring. (3-0/2-1 merged; IxDF, nucamp/MIT Comm Lab, fantasticportfolios, lovable.dev.)

1. Trade-offs explicitly called out = THE architectural/seniority signal;

also attribute your specific role. (3-0; dev.to AWS Community Builder, systemdesignhandbook, DesignGurus, lovable.dev.)

1. Quantification fallbacks: estimate honestly / qualitative before-after /

scope statements. (3-0; sweresume.app, Teal, EduAvenues.)

Refuted claims (do NOT cite)

• "Recruiters spend an average 7.4 seconds initially screening" — vote 1-2.
• "Portfolio first-visit time-on-page is under 15 seconds" — vote 0-3.

Caveats

• Much case-study evidence is from UX/design portfolios — reinterpret visual

stages as architecture diagrams + trade-off analysis for backend work.

• Several sources are career/resume-builder marketing blogs; adequate for

convergent best-practice advice, but none is a rigorous study.

• The 3–5 project band is a curation principle, not a hard number.
• UX-trend specifics (card grid, dark theme) were NOT verified — open.