Backend & AI Systems Engineer
TypeScript / Node.js • Python • Integrações ERP • Sistemas Regulados
I build backend systems and integrations where traceability, consistency and operational continuity are real product requirements.
My work is centered on Node.js, TypeScript and Python, with experience in ERP integrations (SAP/TOTVS), APIs, data flows and regulated environments (BPF/GMP). I am expanding this foundation into AI systems: model integration, MCP, evaluation, guardrails, observability and secure operation.
I work at Laboratório Cristália in the pharmaceutical sector and have maintained PhotoGIMP since 2021.
| Project | What it demonstrates |
|---|---|
| Pimbas | Product backend: domain modeling, authentication, Prisma/Postgres, tests, Docker and operational documentation. |
| MCP Animagine XL | Applied AI engineering: FastMCP, REST API, validation, model management and CPU/GPU execution paths. |
| Astrum | Data-oriented backend: multi-database queries, batched processing and XLSX/JSON reporting. |
| D4Sign Node | Typed Node.js integration for an electronic-signature API. |
| Resulta | TypeScript library for explicit result handling, with tests, migration guidance and releases. |
- Start from the domain and make trade-offs explicit.
- Prefer reproducible systems: tests, documented decisions, containers and clear operational limits.
- Treat AI as a system component that needs evaluation, access control, observability and human review — not as magic.
- Keep sensitive data, internal processes and credentials out of public repositories.
- Reliable AI integrations with MCP, structured outputs and evaluation.
- Backend architecture for regulated and integration-heavy domains.
- Clear technical writing: architecture dossiers, limits and measurable outcomes.
TypeScript · Node.js · Python · NestJS · Fastify · PostgreSQL · Redis · Docker · AWS · GitHub Actions
“Não leio livros sobre o mundo. Saio por aí e vejo por mim mesmo.” — Ezreal