Portfolio: technical SEO case studies with verifiable evidence

Every case study answers three questions in the same order: what needed solving, what was done and how it turned out later. With a date, an instrument and a section on what I can't claim.

Get a free diagnosis For teams and recruiters

A single page application that Google received empty

What needed solving: The site looked complete in the browser, but Googlebot received a title and an empty container.
What was done: Real HTML per route, a Worker in front with real 301s and 404s, and 19 checks before every publish.
How it turned out · Sep 17, 2026: Daily impressions on Google went from single digits in July to between 30 and 66 in September.

JavaScript rendering
Prerender / SSG
301 redirects
Real 404
Sitemap and robots.txt

Read the case: React app on Base44

Putting cardonaseo.com into production without relying on the dashboard

What needed solving: The domain didn't resolve, the platform served crawlers a 350-word template and the sitemap listed noindex pages.
What was done: DNS on Cloudflare, a Worker as custom domain, my own prerender, real 301s and 404s, authoritative sitemap and llms.txt.
How it turned out · Sep 19, 2026: 9,820 requests served in five days with 0 errors, and the audit's 7 discrepancies published with their status.

Domain and DNS migration
301 redirects
Real 404
Prerender / SSG
Sitemap and robots.txt

Read the case: Domain and Worker

Cayu, the SEO tool I designed and verify

What needed solving: Commercial tools gave numbers without a formula: difficulty 0 on a keyword whose top 10 had an average authority of 46.
What was done: Four MCP servers in Python and a React app, generated with Claude Code under my specification and tested with two inputs that differ.
How it turned out · Sep 17, 2026: Three times it showed commercial difficulty was underestimating, two of its own defects found and fixed, and 0.18 USD for 31 keywords in two countries.

SEO automation with AI
Keyword research with real volume
Calculated difficulty (Real-KD)
Structured data audit
Core Web Vitals (median and range)

Read the case: Cayu

How I decided what to sell with data, not intuition

What needed solving: I was going to advertise as an SEO auditor. The biggest search was «agencia seo bogota», and I'm not an agency.
What was done: I measured volumes in four countries, cross-checked two sources where I could and calculated difficulty from the real top 10, in Spanish and in English.
How it turned out · Sep 17, 2026: Positioning decided: technical SEO consultant. Four days after indexing, the home page already ranks 1 for my name; the rest gets measured in October.

Keyword research
Search intent
Clusters and cannibalization
Calculated difficulty (Real-KD)
Research in two languages

Read the case: Keyword research

How to read these cases

A case study without a «what I can't claim» section is advertising. Here every number says where it came from, every chart carries its values in a table, and every case ends with a command so you can check what I claim. The tools I name are linked to their official source, in case you don't know them. The method is explained in the technical SEO blog (in Spanish) and in the guides.

If you're hiring

I work remotely from Bogotá, on US hours, on contract or full time, in English and Spanish. These cases show technical audits, JavaScript rendering, migrations, structured data and AI automation, which is what technical SEO job ads ask for today. What I'm looking for, my stack and how I work with AI is in About, in the section for teams.

My LinkedIn profile