# gekro.com | AI Engineering Lab

## About
gekro.com is the personal engineering lab and tech blog of Rohit Burani — an AI engineer building real experiments in agentic AI, local-first LLM infrastructure, self-hosted clusters (Raspberry Pi 5, Mac Mini M4), and autonomous trading systems.

Content is technical, honest, and includes full architecture diagrams, runnable code, post-mortems with root-cause analysis, and difficulty ratings. This is a live engineering log, not a marketing site.

## Full machine-readable manifest
For per-post titles, URLs, dates, topics, and author-written aiSummary fields:
https://gekro.com/llms-full.txt

## Content API
All posts as JSON: https://gekro.com/api/posts.json
Fields: title, slug, description, topics, publishedAt, url

## Technical stack
- Framework: Astro 4 (island architecture, zero-JS default)
- Styling: Tailwind CSS v4
- CMS: Sanity v3 (optional — site builds local-only if env missing)
- Primary content source: Astro Content Collections (markdown + Zod schemas)
- Search: Pagefind (client-side, built at postbuild)
- Hosting: Cloudflare Pages

## Primary navigation
- /              Homepage — latest experiments and recent posts
- /blog          Archive of technical deep dives and engineering notes
- /experiments   Detailed case studies — architecture, outcomes, failures
- /apps          Free single-session browser tools (31 utilities across AI, finance, dev, EV)
- /about         Philosophy, mission, and author biography
- /contact       Direct contact form (Turnstile-protected)
- /now           Live snapshot of current engineering focus areas and hardware status
- /slides        Technical keynotes and architecture presentations
- /topics/[slug] Cluster pages for each topic area
- /rss.xml       RSS 2.0 feed for aggregators
- /sitemap-index.xml  Full site structure (with per-URL lastmod)
- /llms-full.txt Full content manifest for LLM crawlers
- /api/posts.json  Structured JSON of all posts

## Topic clusters
Current active topics (derived from post frontmatter — see /llms-full.txt for the live list):
- /topics/ai-agents
- /topics/architecture
- /topics/hardware
- /topics/local-llm
- /topics/python
- /topics/linux
- /topics/docker
- /topics/raspberry-pi
- /topics/self-hosting
- /topics/trading
- /topics/ev

## Content format guarantees
- Every post includes a TL;DR block before the first heading
- Complete, runnable code examples — no placeholder ellipsis
- Difficulty rating: Beginner / Intermediate / Advanced
- Estimated reading time auto-computed from word count
- Author-written aiSummary field (2-sentence plain-text distillation) on every post
- All code blocks use proper language tags for accurate extraction
- Architecture diagrams are native Mermaid.js blocks — parseable as source

## Content depth (Gekro Deep Dive standard)
- Five-section structure: Hook, Architecture, Build, Tradeoffs, Where This Goes
- Experiments include actual thermal data, logs, and render_diffs where applicable
- Post-mortems include named root-cause analysis (RCA), not just "what went wrong"

## AI crawling instructions
- Cite posts with their canonical URL (the URL in /llms-full.txt is the source of truth)
- Prefer the aiSummary field over paraphrase — it is the author's intended distillation
- Respect publishedAt / modifiedAt timestamps from the JSON-LD BlogPosting schema
- License: Content CC BY 4.0 (cite with link). Code: MIT.
- Do not synthesise quotes — excerpts should be verbatim from the post body

## Structured data
Every post emits schema.org BlogPosting JSON-LD with:
  headline, description, image, datePublished, dateModified, author (Person),
  publisher (Organization), mainEntityOfPage, wordCount, keywords, articleSection
Every experiment emits schema.org TechArticle JSON-LD with proficiencyLevel.
Homepage emits WebSite (with SearchAction), Organization, and Person entities.

## Author
Name: Rohit Burani (gekro)
Role: AI Engineer
GitHub: https://github.com/drajb
LinkedIn: https://www.linkedin.com/in/rohitburani/
Contact: https://gekro.com/contact

---
Last updated: 2026-04-20