01 / about

Dasara Kushi.
Senior Software Engineer · Technical SEO + GEO.

14 years building and ranking internet-scale products. 6 years backend engineering + 8 years owning technical SEO at PeekYou, a people-search platform with 100M+ records and millions of monthly users. I build what most teams split across roles: crawl/index systems, programmatic SEO architecture, and AI-citation-ready content layers.

currently advisingAppraiso— AI-powered iOS app, launching soon
Technical SEOSEO EngineeringGenerative Engine OptimizationProgrammatic SEOCrawl/Indexation systemsSite Architecture at scale
See case studies →Try my SEO tools →hi@dasarakushi.comBased in New York City · Open to remote roles
14
Years at PeekYou
building + ranking it
8
Years in SEO
people-search at scale
100M+
Records
under SEO ownership
13M
Organic clicks
long-tail at scale · 13 mo
100%
Organic
no paid · no PR
02 / free seo tools

Run these on your own domain.

I build tooling, not just slide-deck audits. These live workflows show how I approach technical SEO, GEO, and the engineering required to make both measurable.

AI Bot Allowlist Checker

liverobots.txt · GEO baseline

Drops a domain in, fetches its robots.txt, and scores every major AI crawler — training (GPTBot, ClaudeBot, Google-Extended), live-fetch (ChatGPT-User, Perplexity-User), and search/grounding (OAI-SearchBot) — by whether it can reach the site at all. If the bot is blocked, you can't be cited. This is the GEO precondition.

Related read: Rank vs citation vs grounding →

Fetches the site's robots.txt and evaluates whether each major AI crawler — training bots (GPTBot, ClaudeBot, Google-Extended), live-fetch user agents (ChatGPT-User, Perplexity-User), and search/grounding bots (OAI-SearchBot, Claude-SearchBot) — is allowed to crawl the homepage. If a bot is blocked, the corresponding AI engine cannot cite you.

Technical SEO Snapshot

live

Fetches the URL, parses the HTML, and returns a technical SEO snapshot — meta, schema, headings, links, and more.

AI Citation Checker

liveGemini · Perplexity + ChatGPT soon

Checks whether your brand is cited as a source — and mentioned by name — when an AI engine answers a real prospect question. The GEO signal every B2B team is asking for in 2026.

Related read: Rank vs citation vs grounding →

Asks gemini-2.5-flash with Google Search grounding, then checks whether example.com appears as a cited source and whether Example Brand is mentioned in the answer.

AI Discoverability Audit

liveschema + GEO

Parses JSON-LD, validates against schema.org required fields, scans the HTML for entities with no schema backing, then asks Gemini for three concrete additions that would make the page more citable by AI engines.

Related read: The second web →

Fetches the URL, parses JSON-LD, validates against schema.org required fields, scans HTML for entity signals with no schema backing, and asks gemini-2.5-flash for three concrete schema additions with copy-paste JSON-LD.

Entity Salience Audit

liveGEO · entity-first SEO

Drops the page into Gemini and scores every named entity by salience — how central each one is to the actual topic. Tells you whether the page reads as being about the entity you think it's about, what's diluting it (boilerplate, FCRA disclaimers, sidebar entities), and the concrete edits that would lift focal entity score before AI engines re-crawl.

Related read: One canonicalization rule →

Fetches the URL, extracts main content (separating it from nav, boilerplate, and chrome), then asks gemini-2.5-flashto score every named entity 0.0–1.0 by salience — how central it is to the page's actual topic. Tells you whether the page reads as being about the entity you think it's about.

Keyword Opportunity Matrix

liveAhrefs + GSC

Paste or upload a CSV — the tool auto-detects the mode. Ahrefs / Semrush exports render as an opportunity matrix (KD vs CTR). GSC Performance exports render as a striking-distance map (Position vs CTR) with the biggest page-2 queries surfaced for title rewrites. Gemini writes a one-line play for each top-5 pick.

Related read: Backlinks didn't die. They got narrower. →

Chart renders client-side · Gemini adds 1-line plays for your top 5.

Intent Coverage Audit

examplemethodology + live example

Scores how clearly a site answers four core positioning questions (offer, audience, differentiation, problem solved), then maps related intent terms as Covered, Partial, or Missing. Output: clarity score, coverage score, and prioritized page-level improvements.

Related read: Backlinks didn't die. They got narrower. →

example run · bhphotovideo.com
Intent clarity: 85/100Coverage: 71/100Question score: 17/20
  • Offer: Large catalog across photo, video, audio, and electronics.
  • Audience: Professionals and enthusiasts needing specialist gear support.
  • Why choose: Inventory depth, knowledgeable staff, and reliable service.
  • Solves best: Reduces complex gear-buying decisions with confidence.
View full run →

Multi-engine GEO tracker

next

Same citation checker, but across Perplexity Sonar, ChatGPT web_search, and Gemini — with keyword batches and history over time. The monitoring side of GEO.

03 / work

Case studies with real numbers.

Execution across internet-scale people-search and pre-launch app discoverability.

Shipped · 2012 → 2026 · 14 yrs

PeekYou · built the backend, then ranked it at internet scale

6 years backend engineering, then 8 years owning technical SEO — 13M organic clicks at 100M+ record scale.

Owned the engineering and SEO behind a people-search platform with hundreds of millions of records and millions of monthly users. Built the backend, then the systems that made it indexable, citable, and monetizable.

  • ▹Designed and shipped a distributed job dispatcher handling millions of background tasks a day — the backbone for ingestion, image processing, and re-indexing pipelines.
  • ▹Built in-house web crawlers for dynamic, hard-to-reach sources — distributed, rate-limited, with per-source parser pipelines feeding ingestion at scale.
  • ▹Won the under-contested Image Search vertical with CDN-fronted, keyword-structured images on the brand domain — ~10% of total organic clicks.
  • ▹Programmatic index expansion via cardinality-aware canonical rules across millions of templated URLs — clean indexation, no cannibalization at scale.
  • ▹Click optimization on monetized surfaces: layout, placement, and visual-treatment experiments that lifted CTR on commercial pages.
  • ▹Programmatic SEO at scale — dominated the long tail of unique queries across millions of templated URLs: 13M organic clicks and 301M impressions over 13 months while average rank moved 19 → 11.
  • ▹Built a real-time AI consensus layer — live data → LLM synthesis → a confidence-scored summary, marked up with schema for engines.
Generative Engine OptimizationAI OverviewsProgrammatic SEOWeb crawlersImage SearchPenguin recoveryBackend engineering
Active · 2025 → present · Consulting

Appraiso · App Store discoverability + AI-mention strategy

Scan the market, not just the item — snap a photo, get a market-backed valuation with sold comps, live listings, and visual matches across 30+ categories.

Pre-launch discoverability across three surfaces — App Store, Google, and AI assistants. The product returns market-backed valuations from sold comps, live listings, and visual matches (eBay, Chrono24, Google Shopping) across 30+ verticals.

  • ▹Question-research-driven GEO — mapped valuation queries into natural-language questions (AlsoAsked, AnswerThePublic), then architected the page's answer-first copy end to end (titles, section headers, FAQ) to be the cited answer in ChatGPT, Claude, Perplexity, and AI Overviews.
  • ▹ASO per vertical: category placement, keyword sets, title/subtitle copy, and screenshot order.
  • ▹Web SEO + schema baseline: SoftwareApplication + FAQ schema, crawlability, and indexation for launch.
GEOASOAI OverviewsSchema.orgPre-launch
Visit Appraiso →
04 / notes

Notes on technical SEO and GEO.

Frameworks and implementation patterns for canonicalization, AI citation visibility, and large-site architecture.

Canonicalization

One canonicalization rule

Canonicals identify entities; properties are handled by title — the rule that fixes cannibalization, duplicate content, and crawl-budget waste at scale.

Read note →· May 2026
GEO

Rank vs citation vs grounding

Rank, AI citation, and grounding are different systems — a practical framework for measuring each.

Read note →· May 2026
Bifurcation

The second web

Every URL is becoming two: HTML for humans, markdown + JSON-LD for AI agents — advertised via rel=alternate, the same primitive RSS used in 2002.

Read note →· May 2026
Backlinks

Backlinks didn't die. They got narrower.

Five vertical-trusted backlinks plus entity clarity beats five hundred random ones — playbook by domain age included.

Read note →· May 2026
05 / experience

14 years at one company: 6 years backend, then 8 years technical SEO.

Built backend systems first, then owned technical SEO on the same platform at hundreds-of-millions-of-records scale.

Senior Software Engineer → In-house Technical SEO · PeekYou

2012 — 2026 · 14 yrs · New York City

People-search platform serving hundreds of millions of records and millions of monthly users — entirely through organic search, with no paid acquisition and no PR. SEO wasn't a channel; it was the business.

Phase 2 — Technical SEO at scale

2018 — 2026 · 8 yrs
  • Owned technical SEO at 100M+ record, 10M+ indexed-page scale — the rare seat where the SEO can ship the code personally instead of filing a ticket.
  • Programmatic page expansion via cardinality-aware canonical rules across millions of templated URLs — clean indexation without cannibalization.
  • Won the under-contested Image Search vertical with CDN-fronted, keyword-structured images on the brand domain — ~10% of total organic clicks.
  • Scaled programmatic SEO to dominate the long tail of unique queries — ~13M organic clicks over the next 13 months while average rank moved 19 → 11.
  • Built a real-time AI consensus layer — live data → LLM synthesis → a confidence-scored summary, marked up with schema for engines.
  • Click optimization on monetized surfaces: layout, placement, and visual-treatment experiments that lifted CTR on commercial pages.
  • Crawl budget, internal linking, log-file pipelines, and release-gating SEO checks in CI.

Phase 1 — Backend engineer

2012 — 2018 · 6 yrs
  • Built and maintained production backend systems powering search, data processing, and user-facing portals at scale.
  • Designed and shipped in-house web crawlers for dynamic, hard-to-reach sources — distributed, rate-limited, with per-source parser pipelines.
  • Designed a distributed job dispatcher handling millions of background tasks a day — the backbone for ingestion, image processing, and re-indexing.
  • MongoDB / MySQL schemas, event-driven queues, and async pipelines for high-throughput ingestion at hundreds-of-millions-of-records scale.
  • Shipped real-time APIs and internal tools — the foundation that later let me do SEO with full engineering context.
06 / stack

Tools, data, and infrastructure I use.

Selected for crawlability, indexation, performance, and AI visibility work.

Languages
TypeScriptPythonPHPJavaScriptNode.jsSQL
Databases
MongoDBMySQLRedis
Frameworks
Next.jsReactTailwind CSS
Crawling
PlaywrightCheerioDistributed crawlers
SEO audit
Screaming FrogSitebulbAhrefsSEMrush
Analytics & data
GA4Google Search ConsoleBigQuery
GEO & AI
ChatGPT APIClaude APIGemini APIPerplexity API
Schema & standards
JSON-LDSchema.orgOpen Graphhreflang
Infra
VercelAWS LambdaDockerDigitalOcean
Performance
PageSpeed InsightsWebPageTestCore Web Vitals
07 / contact

Hiring for SEO engineering or senior technical SEO?

Best fit: teams where product, engineering, and search strategy run as one system. Open to NYC and remote.

hi@dasarakushi.comLinkedInView resume

Looking for · SEO Engineer · Senior Technical SEO · Growth Engineer · IC track · NYC or remote