I build agents
that ship.

I lead the AI engineering practice at GM Mexico, where I’ve shipped LucIA, a hybrid-search RAG co-pilot that cut hallucinations 40% and time-to-insight 60%, and a screen-aware SOP agent that documents processes in real time. I authored the corporate Buy-vs-Build framework now used as the standard lens for new AI initiatives, and I advise the C-suite on where AI actually pays off.

On nights and weekends I design, build, and run Martita — a WhatsApp-native, multi-tenant AI assistant for Mexican SMEs. One codebase, fully isolated per business, where Claude reasons over 21 tools to handle messy, multi-intent Spanish the way a human receptionist would.

~7,500 lines, solo, live with paying customers. It’s also my testbed for production agent patterns before they reach GM.

Before the AI practice I was a data scientist in pricing & incentives, where I shipped Mexico’s first ML incentive-optimization engine — now the standard tool for allocating monthly pricing budgets across vehicle lines.

I started in telecom and distributed systems — core-network engineering at Huawei and network engineering at AT&T across Latin America — before finding my way into data and, eventually, agents.

Along the way: an AI/ML certificate from UC Berkeley Haas, an M.Sc. from Tec de Monterrey, and a B.Sc. in Telecommunications Engineering.