2.5+ yrs
since Jan 2024
Technical SEO Strategist · Kathmandu
Rankings are an engineering problem.
I do the engineering.▋
I'm Kushal Pathak. I fix the code that blocks rankings — and untangle sites that are indexed but not winning: rendering, crawlability, structured data, indexation, and forensic Search Console diagnosis.
50+
brands engineered
10
countries
20+
industries
CPG · SaaS
national brands & funded platforms
// 01 — about
The parts of SEO most people avoid.
Since 2024 I've worked on 50+ websites across 10 countries and 20+ industries — home services, ecommerce, SaaS, health, real estate, manufacturing. My lane is the technical layer: JavaScript rendering, crawl architecture, structured data, indexation strategy, and forensic Google Search Console diagnosis.
I don't stop at recommendations. I build the automation that ships the fix at scale — Apps Script, Python, WordPress plugins, and my own self-built SEO automation assistant. My method is evidence-first: check the data, prove the cause, then fix it.
# education: BIT — Bachelor of Information Technology
2026 →Technical SEO Team Lead / StrategistRankMeTop
2025Technical SEO ExecutiveRankMeTop
2024 JunSEO Trainee
2024 JanSEO Intern
// off the clock — Gosaikunda, Nepal
// 02 — frameworks
Named systems, not vibes.
Repeatable methodologies I run on every engagement — each one built from hundreds of real diagnoses.
FW-01
The Render-Gap Protocol
Reconcile the pre-render HTML payload against the hydrated DOM to expose exactly what crawlers and LLM agents resolve — then close the delta.
FW-02
The Graduation Engine
Classify why an indexed site stays suppressed — impressions without conversions — into discrete failure modes, and compute a data-derived exit path from Search Console.
FW-03
Entity-First Architecture
Build a machine-legible entity ontology and gate every URL through “does this query deserve a page?”, so topical authority compounds instead of diluting.
FW-04
Evidence-Bound Forensics
Interrogate first-party data — Search Console and response headers — before asserting cause. Every finding ships as Issue → Evidence → Solution → Execution.
FW-05
The Machine-Readable Layer
Structured data, semantic HTML, and AI-crawler access engineered so a site is legible to classic search and generative retrieval alike.
// 03 — what i do
Full-stack technical SEO & measurement.
From crawl to conversion tracking — the whole pipeline that turns a page into a ranked, measured, revenue-attributed asset.
01 Technical Audits & Forensics
Page-level, evidence-bound audits with prioritized fixes.
02 Rendering & Crawl Architecture
JS rendering, robots/llms.txt, log & crawl-budget analysis.
03 Schema / Structured Data
Hand-built JSON-LD, entity markup, rich-result engineering.
04 Entity & Topical Architecture
Ontologies, topical maps, query-deserves-a-page planning.
05 Migrations & Indexation
Canonical strategy, product vs. collection indexation, redirects.
06 Core Web Vitals
Field-data-driven performance and rendering optimization.
07 GSC Diagnostics & Recovery
Drops, core & spam updates, the sandbox/graduation problem.
08 Analytics & Measurement
GA4, Google Tag Manager, data-layer & conversion tracking, Looker Studio, BigQuery.
09 Paid-Media Tracking & Feeds
Google & Meta Ads tracking, Google Merchant Center product feeds.
10 SEO Automation & Tooling
Apps Script, Python, WordPress plugins, and a self-built SEO automation assistant.
PLATFORM-AGNOSTIC
WordPress · Shopify · Squarespace · Wix · Webflow · headless & custom CMS
Any platform, any stack — whatever technical issue it surfaces, or whatever the requirement demands, it gets solved. No exceptions.
SELECTED BUILD · TypeScript
SEO Automation Assistant
A self-built bot that automates repetitive technical-SEO and daily ops tasks — the tooling side of how I ship fixes at scale.
// 04 — brands
Work you'd recognize — under NDA.
Client names stay private, but the scale doesn't have to. A sample of the categories I've engineered for.
1.5B+
annual search impressions across managed properties
21M+
annual organic clicks
YMYL
health, CPG & SaaS at national scale
CPG · FOOD
National seafood & BBQ brands
shelf brands · high-volume organic
HEALTH
A top health-authority site
YMYL · six-figure monthly sessions
SAAS
Funded software platforms
IT, CMMS, real-estate tech
ECOM · SERVICES
DTC & multi-geo home services
US · UK · AU · CA · IE
// 05 — skills
Proficiency, self-assessed.
Honest levels — deep where it counts, competent across the measurement stack.
Technical SEO & Crawl Architecture95
Structured Data / Schema Engineering92
JS Rendering & Render-Gap Analysis90
GSC Diagnostics & Data Analysis90
Entity & Topical Architecture88
Analytics & Measurement (GA4 · GTM)85
SEO Automation & Tooling (self-built)84
Paid-Media Tracking & Feeds72
// 06 — the lab
The technical SEO lab.
Four working tools, not screenshots — each running the actual algorithm I use every day, live in your browser.
Google's real matching logic — group selection, longest-match wins, Allow beats Disallow on ties, with * and $ wildcards.
robots.txteditable
testlive
try:
/private/report.pdf as Googlebot vs *, or any path as GPTBot.Entity-grade JSON-LD — @graph wiring by @id, sameAs entity-linking (Wikipedia + Wikidata), areaServed with geo, additionalType (productontology), FAQ hasPart. The structured data an AI can't fake. Fields with | take one item per line.
fieldseditable
json-ld
Google truncates by pixel width, not character count (~600px title, ~920px description on desktop). Measured live with canvas.
inputseditable
google previewdesktop
The Render-Gap Protocol, visualized. Toggle between what a crawler fetches (raw HTML) and what a browser renders after JavaScript. If content only appears after JS, first-wave crawlers can miss it entirely.
viewread-only
Build a reciprocal hreflang cluster for a multi-geo brand (UK / IE / AU). Every alternate page must emit this same full set — or Google ignores it.
localeshreflang | URL — one per line
output
Open Graph + Twitter Card tags with a live link-card preview. (The preview image may not load inside this sandbox — the tags themselves are correct.)
auto-fillpaste a page's <head> HTML
On the live site this becomes “enter URL → auto-fetch” via a small server endpoint.
inputseditable
image
meta tags
"Why isn't my page indexed?" — the exact decision path I walk, from status code down to the Graduation (quality) problem.
Three tiers — pick your level. Easy = fundamentals every SEO should know cold · Medium = day-to-day technical judgment · Hard = the stuff that separates operators. 10 questions each.
// 07 — contact
Have a site that's indexed but not winning?
Send the domain and the symptom. I'll tell you what the data says.
$ emailkushalpathak18@gmail.com
$ phone+977 9840271850
$ linkedin/in/kushal-pathak
$ github/Kushzz18
$ locationKathmandu, Nepal · GMT+5:45