TAYSTUDIO operation notes · tool guides · tech/career · useful posts
![[Career] Samsung Electronics DX G-CS Experienced Hire Interview (AI Vision & Data Optimization)](assets/posts/career-samsung-dx-gcs-2026/cover.jpg)
A first-hand review of Samsung Electronics DX division's G-CS (Global CS Center) experienced-hire process from late 2025 to early 2026: document screening, phone interview, CBT, Work Sampling, and the hiring-manager interview.
![[Admissions] KAIST Grad School Interview (Mechanical Engineering) - Oral Exam Questions & Final Rejection](assets/posts/kaist-grad-mech-interview/cover.webp)
A first-hand review of the 2019 KAIST Mechanical Engineering grad school interview: the real oral-exam questions in the thermo-fluids and dynamics-solids rooms, and how it ended in a final rejection.
![[Admissions] Yonsei Grad School Interview (Mechanical Engineering) - Oral Exam Questions & Timeline](assets/posts/yonsei-grad-mech-interview/cover.webp)
A first-hand review of the Yonsei Mechanical Engineering grad school interview: the application timeline, research plan, and the real oral-exam questions asked.
Move from .github.io to a custom .com in 4 steps. Cloudflare DNS, repo CNAME, Let's Encrypt HTTPS auto-issued, and re-indexing across search engines — UI walkthroughs included.
30 days after migrating to a new .com domain. Indexing gradually OK, impressions follow a spike → dip → recovery pattern. Naver and Cloudflare data carry the load while Google sandbox runs its 6–12 week course.
Naver carries 80% of Korean-site traffic, yet most operators skip Naver SearchAdvisor and let the engine find their site by accident. Registration in 5 minutes, sitemap submission, and manual crawl requests — step by step.
GSC: 2 clicks in 3 months. Naver: 230 clicks in 30 days. Cloudflare: 57 visits a day. Same site, 33–115× gap. Which one is the real traffic, and which one are you missing?
robots.txt is the deal with crawlers. llms.txt is the deal with LLMs. The day chatgpt.com first showed up in Cloudflare referrers, I rewrote llms.txt and explicitly allowed every LLM crawler in robots.txt. An hour's worth of notes.
What Google sandbox is · why a fresh domain disappears from search · the real reason CTR drops · how AI Overview eats traffic. 12 concepts I bumped into the hard way after one month of running a new domain, in one page.
You don't need Lighthouse for diagnosis. curl -w timing variables give you TTFB and total time, sips reads image dimensions and bytes, and Python aggregates per-page asset weight. A 30-line script you can drop into CI for quantitative regression checks.