I build AI systems that ship.

Background

My path to AI engineering wasn't conventional. That's the point.

I started in graphic design and tattoo artistry, where I learned that precision and intentionality aren't optional. Then I spent years writing deep technical analysis: 20+ research reports on decentralized protocol architecture, on-chain data patterns, and economic system design. That work trained me to read complex systems, find failure modes, and communicate findings clearly.

When I discovered LLMs, MCP, and agentic systems, I stopped analyzing other people's infrastructure and started building my own.

I'm Mexican, I operate globally, I speak both En and Es.

Current Focus

Today I specialize in production AI systems engineering. Specifically:

• Custom MCP server design (tools, resources, prompts, elicitations, tasks, MCP-UI mini-apps) with SSE and HTTP Streamable transport
• 3-tier agentic orchestration: lightweight router → domain-specialized agents with native tool_use → Expert Registry with scoped allowlists and live MCP resource injection
• Langfuse observability: every LLM call, tool invocation, model selection, token cost, and streaming event tracked end-to-end with cost attribution per session
• HITL reviewer surfaces: compliance dashboard and EFOS monitor as Next.js interfaces with one-click approval flows wired directly into the agent pipeline
• LLM system testing: 3,960+ passing tests, 91%+ coverage on a solo-built TypeScript codebase
• Full deployment stack: Docker, Railway, CI/CD, rate limiting, security headers, RLS, observability

I currently operate as an independent AI engineer, building DISAI_Conta (a production fiscal AI platform) and taking on select consulting engagements for teams that need real AI infrastructure, not demos.