Blog
README Post Template
home Use this file as a starter when you want to publish a README style project post. What this project does in one sentence. Who it is for. Why it exists. Status: active / prototype / archive…
Cover story
Archive atlas
A living index of essays, briefs, and notebook fragments. Move across the archive to reveal the nearby cluster, then enter a note from its local thread.
Recent essays
From the notebook
Long-form notes, build memos, and working references from the same archive.
Blog
Second Brain Architecture
This note explains the pipeline: 1. Parse frontmatter publish, home, tags. 2. Rewrite Obsidian wikilinks. 3. Render markdown. 4. Generate static HTML pages. See also home and SEO and Performance Guide…
Blog
SEO and Performance Guide
Key points: Canonical URL for each page. Open Graph and Twitter tags. JSON LD structured data. sitemap.xml, robots.txt, rss.xml. Lightweight CSS and static hosting. This page links back to s…
Blog
GitHub to Cloudflare Pipeline
This note documents how markdown updates move to production. blog build compiles notes into static files under dist/. Source: content/posts Static assets: static Outputs: index.html, sitemap.xml…
Blog
Obsidian Linking Habits
Connected writing stays maintainable when links are explicit. note for direct references label for readable anchor text note for deep links into long notes Examples: Start from home Deployme…
Blog
Rust Rendering Notes
This stack uses Askama for page skeletons and Maud for assembled fragments. Askama: predictable layout and metadata templates Maud: ergonomic generation of backlink sections and composed HTML bloc…
Note
Jinhyuk Kim
Jinhyuk Kim holds a B.S. degree from Yonsei University in Korea. He has completed internships in the Mobile Communications Lab and the Robotics and Mobile Networks RAMO Lab and is currently pursuing h…
Papers
Uncertainty-Aware Hybrid Inference with On-Device Small and Remote Large Language Models
Source: https://arxiv.org/pdf/2412.12687.pdfhttps://arxiv.org/pdf/2412.12687.pdf U HLM lets the on device SLM opportunistically skip uplink transmission and server side LLM verification for low unce…