ERICK PANIAGUA · DETROIT
I build AI systems end to end.
A self-taught engineer who takes AI products from prototype to production: voice agents, LLM pipelines, and the eval and safety layers that make them trustworthy. AI-native from day one. I like systems that already have traction and need someone to make them sharper.
THE PATH
Self-taught, by shipping.
I started in e-commerce, fully self-taught: building brands and scaling ad spend across paid social and search. No CS degree, just shipping and measuring.
When COVID grounded air freight, the business was about to break. I rebuilt the logistics layer as a US fulfillment system out of necessity. It became the most valuable asset in the eventual sale of the business to a private equity buyer.
From there I moved into marketing automation engineering, then into building production AI systems independently: a multi-tenant AI voice-agent platform and forward-deployed client systems from 2024, and an AI-native healthcare operations platform from 2025. I own everything technical, end to end.
The throughline: I learn what I need to ship the thing, then I ship it, then I measure it. AI-native from day one.
EXPERIENCE
The path here.
AI Systems Architect (independent), AI-Native Healthcare Operations Platform
2025 to nowA unified AI operations layer for healthcare: voice, chat, and backend agents owning end-to-end workflows. Built product and engineering from the ground up, PHI-first.
Forward Deployed AI Engineer (independent), Voice-AI, Automation & Billing Systems
2024 to nowMulti-tenant AI voice-agent platform plus forward-deployed client systems, built and deployed end to end into client operations. I own everything technical.
Automation engineer, Marketing automation (agency)
2023 to 2024Built marketing automation systems for clients across industries. The bridge from operator to engineer.
Founder, operator, E-commerce business
2017 to 2023Grew a brand from zero to 4.6M dollars in revenue. Rebuilt logistics into a US fulfillment system during COVID, the asset a private equity buyer valued most in the sale.
THE STACK
- TypeScript
- Python
- Next.js
- React
- Node
- LLM pipelines
- RAG
- Voice agents
- Evals
- MCP
- Postgres
- Redis
- Vercel
- Claude Code
- GSAP
This site was designed, built, filmed, and quality-checked by an agentic AI workflow I direct. The pipeline is open source.