About

Lucas Caldas
I'm a software engineer based in Rio de Janeiro. Right now I focus on architecture, mostly on GCP, and on building solutions that put AI to work: agents, MCP, and automations that actually make it to production.
I've spent a good amount of time on backend work with Java, Python, .NET/C#, and TypeScript. Microservices, events, Stripe and PIX payments, messaging, licensing. That background is what helps me treat AI as engineering: clear contracts, observability, and a plan for when things break.
I enjoy studying and shipping agents with ADK, Vertex AI, and MCP. What draws me in is the practical side: tool orchestration, cold start, cost, and reliability. I also keep personal projects going, like Astra, a personalized AI intelligence newsletter.
If you want to chat about architecture, agents, or something you're building, I'm around.
Cloud Architecture / GCP
Cloud Run, BigQuery, IAM, Pub/Sub, and serverless patterns. Infrastructure that scales without becoming an ops headache.
AI Agents & MCP
Multi-agent systems with ADK, Vertex AI, and MCP Toolbox. Integrations with Jira, Looker, and BigQuery, aimed at problems worth solving.
Payment Systems
Experience with Stripe and PIX: event-driven webhooks, mTLS authentication, and subscription lifecycle management.
Backend Engineering
Java, Python, .NET/C#, and TypeScript. Microservices, message queues, and REST APIs built for production.