About Adrian Gaitan — Full Stack & AI Engineer

From Pereira to production AI

I started in software the way most engineers do — chasing the satisfaction of making something work. That curiosity turned into a focus on the hardest, most interesting layer of modern products: putting machine intelligence into production reliably, at scale, where real users depend on it.

Today I work as a Full Stack & AI Engineer at KitchenSync, an AI-first back-office operating system for restaurants. I ship multi-agent orchestration and LLM-powered features into production — the kind of hands-on, real-world AI engineering that separates working systems from impressive prototypes.

How I build

My core belief is simple: in an AI-native product, the model is the orchestration layer, not a feature bolted onto the side. I design systems around that idea — agents that coordinate, recover from failure, and stay observable; event-driven backends that stay correct under load; and interfaces that make all of it feel effortless.

I care as much about the system underneath as the experience on top. A polished UI on an unreliable backend is a liability, and a robust backend with a confusing interface never gets used. I aim for both.

What I work with

On the backend: TypeScript, Node.js, NestJS, PostgreSQL, Redis, BullMQ, and event-driven, microservice architectures. On the frontend: React, Next.js, Astro, and Tailwind CSS. On the AI side: multi-agent orchestration, RAG, vector embeddings, MCP servers, and the OpenAI, Claude, and Gemini APIs.

I’m also building beyond my day job — Beela, an AI-native OS for Colombian small businesses; SwiftControl, my own software venture; and an Agent Memory Layer that gives AI coding agents persistent, structured memory across sessions.

Where I’m headed

I was studying Systems & Computer Engineering at Universidad Tecnológica de Pereira, and I made the deliberate choice to pause the degree. I’d come to believe the role itself is being redefined faster than traditional curricula can follow — AI is changing what engineers build and how, and academia points in a different direction than where the work is actually going. Pausing let me commit to a more concrete path — built around what’s new and what the world actually needs.

That path is to become an agentic software engineer — going deeper into LLMs, fine-tuning, and model design — while shipping world-class products in production. I solve real problems by leveraging the newest frontier tools, learning in the open and proving it with real systems rather than waiting for a syllabus to catch up. I’m drawn to the intersection of distributed systems and machine intelligence, and I’m always open to the next hard one.

Want to build something together?

I’m open to full-time roles, contract work, and interesting problems.

Get in touch