✕André Xavier Bueno

André Xavier Bueno — Belo Horizonte, BR

AI Builder ×
SEO / Analytics

I build the production tools that run and measure marketing — by directing AI coding agents, and verifying everything before it ships.

See the work →

Open to remote / relocation

Method

How I work with AI

01

Decide

I own the architecture and product decisions — evaluating AI-proposed options or reusing proven infra.

02

Direct

I drive AI coding agents (Claude Code) with precise constraints and the project's existing patterns.

03

Verify

I review every change, gate deploys with automated tests, and run 360° security audits.

04

Operate

I deploy, monitor, and use audits + research to keep improving the system.

→ The AI is the speed. The architecture, the marketing expertise and the verification are mine — and I know where AI shouldn't be trusted.

Featured / 01

Early Access
Live site ↗Source — coming soon

Stack

  • Python 3.12
  • Flask
  • SQLAlchemy
  • PostgreSQL
  • Anthropic API
  • Docker
  • Redis
  • Nginx
  • GitHub Actions

DFFCO — Biomedical Evidence Platform

A platform I conceived and built solo, end-to-end, curating the scientific intersection of dehydration, fibromyalgia and cognitive function — each study shown in two registers: clinical and accessible. It started from a personal reason: someone close to me lives with fibromyalgia and cognitive fog.

What I built

  • ✕A 3-source scientific ingestion pipeline (PubMed, Europe PMC, OpenAlex) with cross-source dedup (PMID + DOI).
  • ✕A tested LLM integration (Anthropic Claude API) with prompt-injection mitigation — untrusted study data isolated in <study_data> tags.
  • ✕5-language i18n (~294 strings) and full technical SEO: sitemap, hreflang, JSON-LD.
  • ✕Hardened security (CSRF, rate limiting, CSP/HSTS), 76 automated tests, and a GitHub Actions CI pipeline.
  • ✕Containerized with Docker — web + PostgreSQL + Redis + Nginx.

Engineering maturity

The AI-assisted literature curation is explicitly flagged for mandatory human review before launch — shipping fast with AI without overstating scientific certainty.

Selected work

Six more systems, built solo by directing AI

02In use

Google Tag Manager via AI

Custom MCP · Martech

An AI-agent + MCP server that operates Google Tag Manager programmatically.

  • ✕List, inspect, create and publish tags, triggers and variables by command.
  • ✕Reading config as structured JSON exposes logic bugs the visual UI hides.
  • ✕Recovered broken conversion tracking; audits went from hours to minutes.
  • Claude Code
  • MCP
  • GTM API
  • GA4
  • Google Ads
03In production

Cordoval Tasks

Team/Project SaaS · Bitrix24-style

A project-management platform used daily — built and operated solo.

  • ✕Kanban + SCRUM, real-time chat, CRM, time-clock, monthly PDF reports.
  • ✕405 automated tests gate every deploy; 7 vulnerabilities closed via audit.
  • ✕Integrations: Search Console, GA4, Google Ads, SEMrush.
  • Next.js
  • Fastify
  • Prisma
  • MySQL
  • Socket.io
  • Redis
04In production

Editorial Manager

Multi-tenant editorial/SEO platform

A Python platform that replaced spreadsheet chaos with one source of truth.

  • ✕FastAPI + SQLAlchemy + MySQL; 92 routes, multi-tenant isolation, client portal.
  • ✕Server-rendered HTMX/Alpine frontend (minimal JS); ~9,500 lines of Python.
  • ✕Self-hosted: Nginx, systemd, Let's Encrypt.
  • Python
  • FastAPI
  • SQLAlchemy
  • MySQL
  • HTMX
05In production

Content Studio

Multi-agent content production

A 6-stage agent pipeline that produces editorial content end-to-end.

  • ✕Built by directing Claude Code directly — agent-memory + per-client rules engine.
  • ✕Python automation: Markdown → Google Docs with programmatic formatting.
  • ✕MCP integrations: SEMrush, Canva, Google Drive.
  • Python
  • Claude Code
  • Google APIs
  • MCP
06In production

Editorial Briefing System

Editorial briefing + SEO scoring

A briefing-driven editorial pipeline with an SEO Content Analyzer.

  • ✕Briefing → Content → Review → Publication, with status cascading automatically.
  • ✕Auto-generates Google Drive docs + email notifications; data in Google Sheets.
  • ✕A Content Analyzer scores content against SEO rules and flags issues.
  • Next.js
  • Fastify
  • Prisma
  • MySQL
  • Google APIs
07Implemented · not operated

Institutional Site + AI SEO Feature

Marketing site · headless WordPress

A high-performance site with advanced SEO and an in-product AI feature.

  • ✕Next.js + headless WordPress (GraphQL); server-side JSON-LD, lead capture.
  • ✕An implemented (not yet operated) AI SEO-suggestion feature with prompt-hash caching.
  • ✕Honest framing: built, not yet battle-tested in production.
  • Next.js
  • WordPress
  • Anthropic API
  • Vercel

Capabilities

What I bring

Marketing & Analytics

  • Technical SEO
  • GEO / AEO
  • Google Tag Manager
  • GA4
  • Google Ads / conversion tracking
  • Schema / JSON-LD
  • SEMrush
  • Content strategy

Building with AI

  • AI coding agents (Claude Code)
  • Prompt engineering
  • Model Context Protocol
  • Custom MCP servers
  • Multi-agent pipelines
  • LLM integration (Claude API)

Stack & Engineering

  • Python (FastAPI / Flask)
  • TypeScript / Node
  • React / Next.js
  • HTML / CSS / Tailwind
  • MySQL / PostgreSQL
  • Data modeling
  • REST · JWT · OAuth2
  • Linux · Nginx · Docker
  • Automated testing
  • Audit-driven security

About

An SEO / analytics specialist who builds.

Years of hands-on SEO, analytics and conversion tracking — and I now build the production systems that run and measure that work, by directing AI. I make the architecture decisions, direct the AI, and verify everything (tests, audits, review) before it ships.

My background is unusual on purpose: a B.Sc. in Information Systems plus an M.Sc. and PhD from UFMG — software foundations with doctoral research rigor. I'm looking for AI-builder / marketing-technology / solutions roles, remote or relocation outside Brazil.

Education

  • Ph.D. in Pediatric Dentistry — UFMG
  • M.Sc. in Pediatric Dentistry — UFMG
  • B.Sc. in Information Systems — Centro Universitário UNA

Contact

Let's talk.

Building with AI and want someone who understands the marketing and can build the system that runs it?

Write me →andrexbueno@gmail.com
LinkedIn ↗GitHub ↗Belo Horizonte, Brazil