Innovation Lead — SEO, product & process automation · Limassol

Anton Skryabin builds systems and AI agents that make teams 10x faster.

Six years of hands-on SEO, up to 1.5M+ organic clicks a day. Now I turn that know-how into an in-house platform, AI agents and process changes that do the repeat work for the team.

  • SEO1.5M+daily organic clicks across the portfolio
  • Automation~65 hof manual work a week automated (estimate)
  • Platform800+commits in 4 months — built solo
  • Scale34+domains in 30+ languages

01 — Innovation

Systems that do the work.

I find the work a team repeats every week and give it to software: platforms, agents and better processes. People keep the decisions.

~65 hof manual work automated every week — estimate across 35 automations

SEO OS — the team’s in-house SEO platform

Problem
SEO data lived in 25+ tools. Audits, reports and checks were done by hand.
What I did
Designed and built a Django platform alone: 67 active modules, 50 scheduled jobs, 25+ integrations (Search Console, BigQuery, GA4, CrUX, Bing, Ahrefs, Serpstat, Jira, Slack, LLMs), ~1,800 automated tests. 800+ commits in four months.
Result
A daily technical audit of 11 page types on every locale host (~1,100 headless renders a day). Change tracking of 82,860 pages per run with alerts. A 156 GB local mirror of Search Console data, so any traffic question is answered without API limits.
  • Python
  • Django
  • htmx
  • DuckDB
  • BigQuery
  • Playwright
  • LLM APIs
9detectors run every week over search, analytics, backlink, warehouse and log data

AI SEO strategist agent

Problem
Insights were spread across data sources and chats. Weekly analysis took a person’s time.
What I did
Built an agent with a structured knowledge base of the team’s domains, events and owners. It reads the data, scores opportunities and drafts tasks.
Result
100 scored findings in its first three weeks, a weekly report, a Slack bot and a voice chat.
  • Claude Code
  • Agents
  • GSC
  • GA4
  • Ahrefs
~24Kinternal documents searchable by any AI assistant

RAG knowledge agent + MCP servers

Problem
“How do we do X here?” had its answer somewhere in Jira, Slack, Confluence or Drive — but nobody could find it fast.
What I did
Indexed the documents with embeddings and exposed them, plus plain-English questions over Search Console data, as two MCP servers.
Result
Teammates’ AI assistants answer with sources, and agents query search data without SQL.
  • RAG
  • MCP
  • Embeddings
24locales covered by the AI meta-tag generator

LLMs in production, at low cost

Problem
Writing H1, title and description for thousands of pages in every locale did not scale, and paid APIs were expensive.
What I did
Built an AI meta generator and one LLM classifier reused in 5 workflows. Tested a vendor alternative on 413 cases and rejected it. Moved workloads to self-hosted open models.
Result
Meta tags for every locale on demand. Lower cost, and data stays in-house.
  • Ollama
  • Gemma
  • OpenAI
  • Evaluation
~50%of SEO-risky product tasks caught before development

Process innovation

Problem
Product changes reached production without an SEO review, and SEO ideas had no clear way into the roadmap.
What I did
Wrote SEO involvement guidelines for product work and added an “SEO related” field in Jira that adds the SEO team automatically. Started SEO review in grooming and an innovation intake with RICE scoring.
Result
SEO is in the loop by default, and automation ideas are ranked by impact, not by who asks loudest.
  • Jira
  • Confluence
  • RICE

02 — SEO

SEO at scale.

Technical, international and generative-search SEO for a portfolio of high-traffic consumer products in 30+ languages — the foundation the tools are built on.

84%of ~8M verified Googlebot hits went to API endpoints, not pages

Googlebot crawl waste, found in server logs

Problem
Nobody knew where Googlebot actually spent its crawl budget.
What I did
Analysed ~8M Googlebot requests from server logs, verified real bots by IP and grouped URLs by shape. Found that most of the crawl went to API endpoints and that ~30K hits came from fake Googlebots.
Result
An “always server-side render for Googlebot” fix, and a permanent crawl monitor (~1M verified fetches a day) inside SEO OS.
  • Log analysis
  • Crawl budget
  • Rendering
  • Elasticsearch
2×clicks for a locale after moving it to its own subdomain

International: locales on subdomains

What I did
Wrote the canonical and hreflang requirements for moving two locales to new subdomains, then tracked the result.
Result
One locale doubled its organic clicks. New Bulgarian and Hebrew locales went live in 2026.
  • hreflang
  • Canonicals
  • International SEO
+700clicks a day from one indexation fix

Search pages: indexation and canonicals

What I did
Got ~300 more search pages indexed and wrote a new title template for search pages. Later wrote the URL-normalisation and canonical rules for search.
Result
About +700 organic clicks a day from the newly indexed pages.
  • Indexation
  • Templates
2 daysuntil traffic moved to the new domain

Domain migration without a drop

What I did
Specified and ran the move of a satellite domain: one-hop page-level 301s, a nameserver switch to Cloudflare and a Change of Address in Search Console.
Result
Search traffic showed up on the new domain two days after go-live.
  • Migration
  • 301
  • DNS
1–3Klook-alike domains checked every day

Brand protection and DMCA

What I did
Owned a page-drop checker bot, triaged DMCA and “shadow DMCA” drops and moved pages back through the de-DMCA flow. Found brand-abuse domains and got them transferred to the company.
Result
Page drops are handled in one clear process, and copycat domains are caught early.
  • DMCA
  • Brand
  • Domains
  • 30M+monthly organic visits on the flagship
  • Top-600global Semrush rank of the flagship
  • 0 → 10Kdaily clicks for satellite projects
  • CWVsingle owner of Core Web Vitals regressions via CrUX

03 — Experience

Experience.

From specialist to team lead to innovation lead — mostly at one company, with growing scope.

SEO Innovation Lead

Mayflower · Sep 2026 – now

  • Own the SEO automation roadmap: internal tools, data pipelines and AI systems for SEO and Growth.
  • Lead the SEO Innovations team, including Generative Engine Optimization (visibility in AI answer engines).
  • Run the innovation intake with RICE scoring and drive adoption of automation in day-to-day work.

SEO Team Lead

Mayflower · Jul 2025 – Sep 2026

  • Owned SEO strategy for several high-traffic international products across Tier 0–3 markets; monthly SEO review with CEO, CPO and CGO.
  • Managed and mentored the on-page SEO team: prioritization, QA, performance tracking.
  • Acted as SEO Product Owner with engineering, design, content, analytics and legal.
  • Built SEO OS (see Innovation).

SEO Specialist → Senior

Mayflower · Jul 2023 – Jun 2025

  • Owned title, H1 and meta templates and the SEO requirements in product tasks.
  • Set up white-label domains and their DNS; ran DMCA triage and brand protection.
  • Technical audits, Python automations and Looker Studio reporting.

SEO Specialist

IQ Option Global · Jan 2021 – Jun 2023

Technical and on-page SEO for an international fintech product.

SEO

Technical SEO, crawl budget, log analysis, rendering, Core Web Vitals, international SEO, migrations, GEO.

Engineering & AI

Python, Django, SQL, BigQuery, DuckDB, LLM APIs, agents, RAG, MCP, self-hosted models.

Leadership

Team lead and mentor, SEO product owner, process design, work with C-level.

Education & certifications

  • ITMO UniversityMTech, Neurotechnologies and software engineering · 2019–2021
  • North-Eastern Federal UniversityBTech, Informatics and computer engineering · 2015–2019
  • LinkedIn LearningLeadership Transitions · Influence Skills for Leaders · Strategic Thinking

04 — Contact

Have a process that should run itself? Let’s talk.