Backend Software Engineer · 7+ years · Python / FastAPI I build the infrastructure AI agents run on — MCP, OAuth 2.1, observability.

I design and ship backend systems that stay reliable in production — clean integrations, solid data flows, and observability so they keep running as requirements evolve. Last few years building a regulated digital-banking platform; earlier, geospatial systems at scale.

A secure, production-grade gateway that gives AI agents audited, least-privilege access to Google Drive and Slack over the Model Context Protocol — without ever leaking the user's identity token to a third party (Confused Deputy prevention, proven by a regression test).

FastAPI · MCP / Streamable HTTP · OAuth 2.1 (RFC 9728 / 7591) · Redis · OpenTelemetry Built with an AI-native SDLC (Claude Code + MCP, spec-driven), with CI, security scanning, and a full test suite.

➡️ github.com/XxTeosxX/mcp-agent-gateway

• Languages: Python, TypeScript, Bash · SQL
• Backend: FastAPI, Django/DRF, REST, GraphQL, microservices, BFF
• AI infra: Model Context Protocol (MCP), tool orchestration, OAuth 2.1 trust boundaries
• Data: PostgreSQL / PostGIS, Redis (Streams, Lua, rate limiting)
• Cloud / Ops: AWS (EC2, RDS, Lambda), GCP (Cloud Run), Terraform, Docker, CI/CD, OpenTelemetry

• Diletta Solutions — Backend Developer / Tech Lead · digital banking (regulated), BaaS integration, GCP
• BlincTek (OneBlinc) — Backend Developer · fintech, Django/DRF, Plaid, data validation
• Oowlish — Backend Developer · geospatial platform (PostGIS), AWS + Terraform, US clients

📍 Fortaleza, CE — Brazil · Remote · 🇧🇷 Portuguese (native) · 🇺🇸 English (professional working)

_{Counts code across pinned + owned repos. html,css hidden so the bar reflects engineering work, not page markup.}