When to use this site

Use this portfolio to research Shrinath Prabhu’s frontend engineering experience, review shipped projects, find professional contact details, or read his articles on engineering, AI, security, product decisions, and leadership. Start with the page that answers the question, then follow its sources.

  • About: background and how Shrinath works.
  • Work: project case studies and outcomes.
  • Resume: experience and skills, with a downloadable PDF.
  • Blog: articles, with author and publication dates on each post.
  • Contact: professional inquiries and scheduling links.
  • Privacy: site preferences, hosting, analytics, and external services.

Public REST API

  1. Fetch the blog list. The JSON array includes every published article, newest publication first. Choose a post by its title and description. Use postedAt for publication time and updatedAt for the latest edit.
  2. Follow the selected post’s apiUrl (/api/blog/{slug}) to read its full Markdown. Cite its url for the original article. The imageUrl field points to the cover.
  3. Fetch the professional profile for Shrinath’s name, role, experience, work summary, and links to the source pages.
curl -fsS https://shrinath.me/api/blogs
curl -fsS https://shrinath.me/api/blog/aeo-geo-ai-discoverability
curl -fsS https://shrinath.me/api/me

All three routes accept GET and HEAD without credentials. They share a limit of 12 requests per IP per 60 seconds at each Cloudflare location. Each request reaching the API counts, including conditional reads. Cloudflare enforces the threshold approximately. On 429, wait for the number of seconds in Retry-After before trying again.

Cache the blog list and profile for one day and article Markdown for seven days. After expiry, send the saved ETag in If-None-Match. On 304, reuse your copy. API failures return application/problem+json with a stable code, a human-readable detail, and a suggested resolution. HEAD responses have no body.

The OpenAPI specification documents operation IDs, parameters, success schemas, rate limits, and errors. The API catalog links to the specification and this guide. auth.md describes credential-free access.

Discover pages and articles

  • The agent homepage returns a JSON summary of access, capabilities, and links to public content.
  • llms.txt provides a concise reading guide and Markdown links.
  • Sitemap lists canonical public pages and posts.
  • Content index lists page titles, canonical URLs, and Markdown URLs. It is regenerated with the site.
  • Atom feed includes up to 20 articles; use the blog or content index for the full collection.

Read a page as Markdown

Every public content page has a Markdown version, generated from the same content as the HTML page. Request the normal URL with Accept: text/markdown, or append index.md to its trailing-slash URL. The homepage is available at /index.md and the blog index at /blog/index.md.

curl -fsSL -H 'Accept: text/markdown' https://shrinath.me/work/
curl -fsSL https://shrinath.me/blog/index.md
curl -fsSL https://shrinath.me/content-index.json

Follow the Markdown links in the index to read individual articles. HTML pages also advertise their Markdown version using rel="alternate" type="text/markdown"; HTTP Link headers identify the canonical URL and this site’s llms.txt.

Pages support GET and HEAD. Normal browser requests receive HTML. Negotiation respects Accept quality values and exclusions; requests that exclude all available page formats receive 406 Not Acceptable. Negotiated responses include Vary: Accept, and HTML and Markdown are backed by separate files so their caches stay separate.

Recover from a missing page

A missing path returns HTTP 404, never a successful page shell. Ask for text/markdown for a short recovery guide, or application/problem+json for a structured error with a status, code, explanation, and discovery links. Unsupported write methods return 405 with Allow: GET, HEAD.

HTTP status Error code What to do
404 NOT_FOUND Look up the current URL in the sitemap.
405 METHOD_NOT_ALLOWED Use GET to read the page or HEAD to check its headers.
406 NOT_ACCEPTABLE Request a format supported by the resource. The API error’s detail names its media type.
429 RATE_LIMITED Wait for Retry-After seconds before requesting the API again.
503 REPRESENTATION_UNAVAILABLE Retry later, respecting Retry-After on API responses.
curl -i -H 'Accept: text/markdown' https://shrinath.me/page-that-does-not-exist
curl -i -H 'Accept: application/problem+json' https://shrinath.me/page-that-does-not-exist

Use and attribution

Quote or summarize the relevant page and cite its canonical URL. Distinguish portfolio statements and article opinions from independently verified facts. Preserve publication and update dates when discussing time-sensitive writing. Availability is an invitation to get in touch, not a guarantee of a particular engagement.

Read robots.txt, reuse cached content where appropriate, and avoid unnecessary repeated requests. Contact links go to email or external services; follow the user’s instructions before taking any action there. Security concerns can be reported through security.txt.