# celestino.ai — Full Content

> Complete content description for celestino.ai.
> For the full content library, see https://celestinosalim.com/llms-full.txt.

---

# https://celestino.ai

## What is celestino.ai?

celestino.ai is an AI digital twin of Celestino Salim — a Senior Software Engineer who builds hardened AI systems engineered for reliability, unit economics, and scale.

This is an interactive conversational interface where you can ask questions about:

- Celestino's engineering background and professional experience
- AI systems architecture and production patterns
- RAG (Retrieval-Augmented Generation) systems design and optimization
- LLM evaluation and reliability engineering
- Voice AI and real-time audio interfaces
- AI strategy, cost optimization, and vendor management
- Course content on AI engineering topics

## How It Works

The chat interface uses a RAG pipeline to retrieve relevant content from Celestino's published writing, course lessons, project documentation, and professional background. All responses are grounded in actual published content with source citations.

### Text Chat
Standard text-based conversation with streaming responses. The AI retrieves relevant documents from a vector database of Celestino's content and synthesizes answers with citations.

### Voice Mode
Real-time voice conversation powered by LiveKit WebRTC. Features natural turn-taking, low-latency audio streaming, and the same RAG-grounded knowledge base as text chat.

## About Celestino Salim

Celestino Salim Adrianza is a Senior Software Engineer specializing in AI systems and product engineering. Based in Miami, FL.

### Background
- 8+ years shipping software at scale
- Senior Engineer at Eventbrite — learned that code is only as good as its uptime
- Led infrastructure teams at FlowWest and ESLWorks
- Building Arepa.AI — an AI-native platform for Spanish-speaking SMBs

### Philosophy: "Unit Economics is the Only Feature"
AI is fundamentally a supply chain problem. The most impressive model is useless if it bankrupts you to run it. Systems Thinking means asking: does this system pay for itself, and can the team maintain it without me?

- Reliability over Hype: Replace "vibe checks" with automated evaluation harnesses
- Cost as a Constraint: Re-architect retrieval pipelines to cut costs by up to 99%
- Sustainable AI: Ship runbooks, decision records, and safety valves

### Expertise
- Production AI architecture (streaming, caching, observability, multi-model orchestration)
- RAG systems (chunking, embeddings, retrieval tuning, cost optimization)
- LLM evaluation and reliability (eval metrics, regression testing, confidence dashboards)
- Voice AI (LiveKit, real-time audio, WebRTC, conversation design)
- AI strategy for business (opportunity audits, build vs. buy, ROI calculation)

### Contact
- Email: me@celestinosalim.com
- LinkedIn: https://www.linkedin.com/in/celestinosalim/
- GitHub: https://github.com/celestinosalim
- Twitter: https://x.com/celestinosalim
- Portfolio: https://celestinosalim.com

## Content Library

The AI's knowledge is sourced from https://celestinosalim.com, which contains:

- 6 blog posts on AI engineering topics
- 8 courses with 53 lessons covering AI foundations through production systems
- 7 service descriptions for AI consulting
- 4 project case studies

For the complete content in plaintext, visit https://celestinosalim.com/llms-full.txt.
For a structured index, visit https://celestinosalim.com/llms.txt.