AI Product Engineer · Senior Full-Stack Engineer · Builder · Founder @ iTOP
I turn ambiguous user pain into shipped software, using AI as leverage.
I build software users love to use.
I'm an AI Product Engineer from Brazil with 12+ years building production systems across healthcare, fintech, Web3, event platforms, and administrative systems.
My strongest area is applied product engineering: understanding ambiguous user problems, designing the workflow, building the full-stack solution, and shipping it to production. I use modern engineering and AI tools to solve real business problems end to end.
Today I work through agent orchestration: Claude Code, Cursor, Codex, custom skills, structured specs, and harnesses that coordinate agents to build with speed, engineering discipline, and quality. The goal is not to make demos. The goal is to ship software that keeps working after real users show up.
- 🔭 Working on: AI-powered SaaS, agentic workflows, RAG systems, and full-stack products
- 💬 Ask me about: building software users love to use, AI product engineering, full-stack architecture, and agent-orchestrated development
- 🌍 Open to: AI Product Engineer, Forward-Deployed Engineer, Solutions Engineer, Founding Engineer, and Senior/Lead Full-Stack roles
| Product / project | What I did | Proof |
|---|---|---|
| iTOP | Founded and built a multi-tenant event-registration SaaS with payments, QR check-in, WhatsApp flows, terms signature, analytics, and AI-first support workflows. | 15K+ registrations, R$5M+ processed, lean team of three |
| SwitchCare | Built payroll automation for a US healthcare staffing platform and led the move from Node + Angular + Flutter to a unified TypeScript stack. | 10K+ users, one-click daily payouts replacing manual Excel payroll |
| Unicrow SDK | Built the official TypeScript SDK and production indexer for a decentralized escrow protocol on Arbitrum and Base. | Open-source SDK, thousands of on-chain transactions indexed |
| WAX NFT / OnGaia | Built Web3 marketplace features, minting flows, auctions, real-time status UI, media optimization, and live-streaming sales. | Production NFT marketplace work across WAX and EVM ecosystems |
| tgmarinhopro.com | Maintain a personal technical blog and portfolio where I document what I build and learn. | 14+ years writing, 100+ public blog entries |
More projects, writing, public code, and proof links in the full portfolio.
| Repository | Stars | What it is |
|---|---|---|
| itop-lgnd |  |
Full-stack SaaS for large-scale events: registration, payments, check-in, and analytics |
| tgmarinho-ai-blog |  |
Bilingual personal website and technical blog built with Next.js, TypeScript, and Velite |
| pi-tg |  |
Open-source AI coding agent toolkit and TypeScript monorepo |
| canetaco |  |
Brazilian electronic signature SaaS with WhatsApp workflows |
More repositories and proof links in the full portfolio.
Auto-generated daily by lowlighter/metrics — static SVG, no flaky third-party calls.
- Customer-first engineering: start from the pain, not the framework.
- Spec-driven execution: turn discovery into clear specs, workflows, and acceptance criteria.
- Agent-orchestrated delivery: use Claude Code, Cursor, Codex, MCP, custom skills, and harnesses to coordinate agents, engineer solutions, and ship faster with quality.
- Full-stack ownership: architecture, frontend, backend, mobile, data, payments, deployment, and production support.
- Pragmatic product sense: ship the useful version, measure what changed, and improve from real usage.
- LLM integration in real products
- RAG pipelines and grounded assistant experiences
- Tool use, MCP, and agentic workflows
- Evaluation loops, guardrails, and observability
- AI-first internal tools and support automation
- Harness engineering for reliable agent-orchestrated development
Product engineering
Data, infra, and AI
Web3
I have worked across support, government systems, freelancing, technical writing, international consulting, Web3, healthcare, and now my own SaaS product.
The pattern is consistent: I enter messy operational context, understand the people and constraints, design the system, write the code, and ship something useful.
I am interested in AI product engineering, applied AI, developer tools, and product-led teams where customer context sits close to engineering.
🔗 tgmarinhopro.com · 💻 github.com/tgmarinho · 💼 linkedin.com/in/tgmarinho