Osvaldo (Waldo) Restrepo
Full-Stack & Applied AI Engineer — building LLMs, agents, and voice systems wired into real ops.
Boquete, Chiriquí, Panama · bugfreewaldo@gmail.com
Skills
A quick snapshot of what I use most.
AI / ML
- LLMs
- RAG
- Agents
- Whisper / TTS
- Prompt design
- Evals
Platform
- Node / NestJS
- Python / FastAPI
- Postgres / MongoDB
- Redis
Frontend
- Next.js
- React
- TailwindCSS
- TypeScript
Proof of Skills
Concrete examples and metrics. Replace the placeholders with your real numbers.
RAG & LLMs
- RAG over hospital policies (LangChain) with JSON/tool calling
- p95 latency ~900ms on Pinecone small index (replace with your real metric)
- Role-based access + audit-friendly logs
Agents & Voice
- Asterisk/FreePBX + Twilio SIP trunk; intent-routing to tools
- Call summaries + CRM notes via server actions
- Human handoff with transcript
Automation & Ops
- Stripe → QuickBooks payout reconciliation (Zapier)
- Single expense per payout; fewer reconciliation errors
- Hours saved monthly (replace with your real number)
Tip: make each bullet measurable (p95 latency, accuracy delta, hours saved, $ saved).