AI Engineer · LLM Integration · Business Automation · Cloud Infrastructure
Bilingual Electronics Engineer (Spanish/English C2) with 5+ years of experience in IT consulting, corporate IT infrastructure, scientific research, intelligent transportation, and banking, in both employee and independent roles across Latin America. That multi-domain background shaped a systems-first approach: I think in terms of architecture, from the data layer to the business layer, and design accordingly, always weighing what a decision costs or saves the business, not just whether it works technically.
💼 What I focus on
- LLM Integration: multi-model AI platforms, self-hosted and API-based, taken from prototype to cloud deployment and validated end-to-end.
- Automation: workflow orchestration and reporting pipelines built to fail safely, with automatic validation and retries instead of silent errors.
- Cloud & DevOps: containerized deployment and infrastructure as code, with every component validated on deployment, not assumed to work.
- AI-assisted development: I use native AI tools (Claude Code) as part of my daily development workflow.
🛠️ How I work
- Architecture in writing before code: scoped, documented decisions so any team member can maintain the system, not just me.
- LLM output is untrusted until proven otherwise: generated content passes deterministic checks — format assertions, schema validation, retries with backoff — before anything ships.
- Pipelines fail closed, not open: health checks block a pipeline from advancing on integrity failures — row counts, referential checks, access rules — instead of letting bad data reach a dashboard.
- Reproducible by default: containerized, version-pinned configs that behave identically across dev, staging, and production — no "works on my machine."
- Honest scope: explicit about what's a prototype and what's in production, without claiming results I can't back up.
📍 Caracas, Venezuela (UTC-4, full overlap with US business hours) · Open to independent remote, hybrid or on-site work.
| Project | Description |
|---|---|
| salesreport-automation | Self-hosted reporting pipeline — n8n orchestration, PostgreSQL analytics, AI narrative generation, deterministic PDF rendering with retries and backoff. |
| ai-multimodel-chatbot | Self-hosted multi-model AI chatbot on GCP — OpenWebUI behind hardened NGINX/Docker, pay-per-token access to OpenAI, Anthropic, xAI and DeepInfra via OpenRouter. |
| insurance-analytics | End-to-end data pipeline for a synthetic insurance portfolio — star schema, row-level security on Supabase, four Looker Studio dashboards. |
| goldenmovies-webapp | Bilingual cinema booking experience — Next.js 16, React 19 Server Components, interactive seat map, typed API layer with rate limiting. |
| agrotech-controlsystem | Local-first IoT architecture for hydroponic control — six C4 diagrams, FSR/SAD/ADRs, an ICS/SCADA-style one-way trust boundary to the cloud. |