Hi, I'm Juan G. Carmona
Software & AI Architect • Secure Systems • Cloud & DevOps
I’m a software & AI architect focused on secure, scalable platforms and agentic AI systems. I design and lead full-stack solutions with clear boundaries (runtime vs persistence), strong DevSecOps, and privacy-first AI: LLM orchestration, RAG pipelines, MCP-style tool integrations, and observability from day one.
- AI Architecture: Agentic runtimes, OpenAI-compatible APIs, streaming events, tool use, retrieval.
- Private AI & RAG: Document pipelines, markdown casting, vector search (ChromaDB), grounded responses.
- DevOps & Infrastructure: Azure, Docker, Terraform, GitHub Actions, CI/CD
- Cybersecurity: Secure-by-design practices
- Clean Code: Domain-driven design, clean architecture
- Mentorship: Coaching developers, engineering teams, startup founders, and tech leaders
I help companies, from startups to enterprises, ship reliable, secure, and cost-aware software. Hands-on or leading teams, I bring calm execution, technical excellence, and a bias for outcomes.
👥 How I Work
- Architecture & Delivery: I design for scalability, security, and developer experience—always with business results in mind.
- Secure by default: identity, least privilege, secrets management, telemetry hardening.
- AI with guardrails: deterministic event flow, monitored tool use, reproducible retrieval.
- Mentorship: I form and coach teams that integrate seamlessly into client organizations.
- Community: I share knowledge, build open source, and support fellow engineers.
Some of my recent Projects
- Sentra Brain (org): Private, modular AI platform (LLM orchestration, RAG pipelines, agentic runtime, FastAPI gateway, React frontends, Microsfot Agent Framework). Built for privacy-first deployments. https://sentrabrain.com/
- Star Wars Death Star arc42 Documentation: Inspired by the Star Wars universe, this project provides an educational example of applying arc42, a popular framework for structured software architecture documentation, to a complex system, exploring each section in detail.
- Mermaid Playground: This repository contains various examples and guides on how to create and represent different types of UML and C4 diagrams using Mermaid.
- FastAPI REST API Tutorial: This repository is an example of a RESTful API implemented with FastAPI. It includes concepts such as** DDD, Vertical Slices, Dependency Injection, Security, Swagger**, and more. NOTE: README.ms is in Spanish.
- Hunting The Hunter (episode one): Real-world forensic analysis of developer-targeted malware disguised as a job offer.
- Lyra: Lightweight CLI tool to efficiently download full videos or audio from YouTube videos.
🏔️ Fun Fact
Outside of tech, I’m an alpinist, climber, skydiver, and ski mountaineer. I like to push limits, both on the ground and at 3,000 meters.
